16 May, 2001

Attendees of the

Aquacultural Waste Management SymposiumJuly 22-24, 2001

Dear Attendees:

Attached are two articles I have written regarding my research on nitrification

and the nitrif~~g bacteria in culture systems. One is a summary of two

papers I have published on the phylogenetics of nit~kg bacteria, The .

original papers are copyrighted by the American Society for Mcrobiology but

can be freely accessed at their web site www.asm.org! under Journals and.

using the search term "Hovanec".

The second article is my presentation given at the WorM Aquaculture Society

meeting held in Orlando Florida this past January.

I hope you find. these useful. If you have questions I can be reached. athovanec.marineland.corn SincerelyTimothy A. Hovanec, Ph.D.Chief Science OKeer

6100 CONDOR DRIVE ~ MOORPARK, CA 93021 ~ 806! 529-1111 ~ FAX 805! 5293030

0

!

Linking Nitrification and the Nitrifying Bacteria

TIMOTHY A. HOVANEC, Ph.D. 1

Chief Science Ofhcer 1, The Aquaria Group - Aquaria Aquatic Research Laboratories, 6100 Condor Or.,Moorpark, CA 93021, 805! 553-4446, Fax 805! 529-0170, hovanecOmarineland.corn

For over 3 decades, the bacteria deemed responsible for nitrification, or biologicalfiltration, in closed aquatic systeins were thought to be Nitrosomonas europaea andNitrobacter winogradskyi. Recent studies using modern methods of molecular biology,including cloning, DNA sequencing, DNA fingerprinting, and fluorescent in situ hybridiza-tion FISH!, have deinonstrated that this is not the case. The data show that there is signifi-cant diversity of ammonia- and nitrite-oxidizing bacteria. The bacteria responsible forammonia oxidation in freshwater and saltwater systems are newly discovered members ofthe Nitrosospira and Nitrosomonas genera. Nitrite-oxidation in freshwater and saltwatersysteins is performed by bacteria belonging to the phylum Nitrospira, which are not closelyrelated to Nitrobacter winogradskyi.

INTRODUCTION

O Timothy A. Hovanec 2000 Page 1

Nitrification in closed aquatic systems, whetherthey are 40 L or 40,000 L in volume, is one of, if notthe inost critical aspects of water quality management.Nitrification is defined as "the oxidation of ammonia to

nitrite, followed by the oxidation of nitrite to nitrate, bybacteria". In closed aquatic systems, nitrification is themost efficient method to remove ammonia from the

culture water before the ammonia reaches a toxic con-

centration. The source of ammonia in the culture water

is the aquatic organisms themselves. Ammonia is theprincipal nitrogenous waste product of teleosts andsome invertebrates, Therefore, if not controlled, the

ammonia concentration in a closed aquatic system willcontinually increase. Aminonia is also toxic to aquaticorganisms resulting in central nervous system impair-ment and eventual death. Acute toxicity of ammonia atconcentrations of O,S to 1,0 mg/L NH3-N! have beendemonstrated. Chronic low levels of ammonia have

also been shown to retard fish growth.

Therefore, a mechanism to rid the culture water of

ammonia on a continuous basis is necessary for thesurvival and development of aquatic species main-tained in closed aquatic systems,

For over 100 years, it has been known that certaingroups of bacteria oxidize ammonia to nitrite andnitrite to nitrate. These bacteria are dassified as

belonging to the family Nitrobacteraceae but it is real-ized that the members of this bacterial family are notall phylogenetically related �!.

The most commonly studied ammonia-oxidizingbacteriuin AOB! is ¹trosonionas europaea whileNitrobacter winogradskyi is the most frequentlystudied nitrite-oxidizing bacterium NOB!, However,other species have been known and isolated from sea-

water, cooling towers and other aquatic environments 9, 11!,

In the field of aquatic filtration, which for thepurposes of this review includes wastewater treatment,aquaculture, public aquaria and the ornamental fishhobby, it has been readily accepted that ¹trosomonaseuropaea and Nitrobacter winogradskyi are the princi-pal AOB and NOB, respectively, responsible for nitrifi-cation. Many scientific papers, books and popular arti-cles have been written about these bacteria and their

important role in water filtration,

AQUATIC FILTRATION SYSTEMS

The three major components of a closed aquaticfiltration system are: mechanical, chemical and biologi-cal. Mechanical filtration is the reinoval of particulatematerial from the water by some type of straining. Thestraining material can be screens, foam pads, spongesor fibrous cartridges to naine but a few!, In all cases,they act on particles in the water and not dissolvedsubstances.

Dissolved substances, such as phenols and tannins,which discolor the water, are removed chemically. Themost common chemical filtration medium is activated

carbon

While mechanical and chemical filtration are

important, almost all aquatic organisms can live inwater that is turbid and/or discolored. The same cannot

be said for water with a high ammonia concentration.

This is why biological fiitration is the most import-ant component of the filtration system. Much work hasbeen done engineering many types of media and kindsof filters on which to grow the nitrifying bacteria, Theengineering has also included calculating water flow

alpha subdivisiondelta subdivisionRhodcpseudcv nones pal ustri�

I'trobac sr vrinottrads t

/gamma subdivisionr

Nttrosp ra group

Magnetobacterium Oavaricumt espy IliumPenoo rictans

Mtrospira sp. --~M resp ra marina M ccoccvs mob is

hlftrosocoocus oceanus~ nfcbacfttus tsrrootridans

l4 osospira brisndMtiosovritrio tenuis

Mtrosostrira mullitorrnlaM ~as V rasa>.

Mtroscmonas crf otofsrans /MtrasamCnaa marintr'

la terrugfnea

molliesL~Ni hesanXnas nal opt I afvarroscmonas eut opna

Mtrcsomonas Aurcpaea

beta subdivision

FIG, 1, Phylogenetic relationships of ammonia- and nitrite-oxidizing bacteria.

Page 2I Timothy A. Hovanec 2000

rates, water retention times, and various other parame-ters necessary for building a filtration system. Further,there are many studies on the rates of nitrification andother physiological aspects of the nitrifying bacteria inclosed aquatic systems.

There have also been studies on determining thebacteria themselves but these studies were misleadingat best because it has been determined that

Nitrosomonas europaea will generally out-competeother ammonia-oxidizing bacteria when one is attempt-ing to obtain pure cultures of these bacteria.

This was the general state of the accepted knowl-edge in the field of aquatic filtration in the early 1990'swhen I started work on my Ph;D, dissertation at theUniversity of California, Santa Barbara.

PHENOTYPIC VERSUS PHYLOGENETIC

CLASSIFICATION

To further understand why prior studies weremisleading one needs to realize that while phenotypicclassification of microorganisins is useful for distin-guishing closely related cultivable species, it is unsatis-factory for interrelating distantly related taxa anduncultivable forms see �2! and references therein!,The traditional classification scheme for microorgan-isms is based on tests of physiological responses andbiochemical reactions �!. Thus, in order for a novelorganism to be classified, a pure culture of theinicroorganisin in question is needed so the requisitetests can be performed. The accompanying responses,it is assumed, give a clue as to the genetic relationshipof the organism to other microorganisms,

While it is clearly not impossible to cultivatenovel organisms, and no organism can be considered

impossible to culture in the strictest sense, many organ-isms resist cultivation. The lack of culturability leadsto an underestimation of microbial diversity in naturalsamples. Further, a bias in cultivation can give undueiinportance to bacteria which, because of their ability togrow in a pure culture versus other members of theassemblage, play a seemingly important role in theenvironment. However, in the natural setting the culti-vable bacteria may be of minor iinportance, relative tothe uncultivated types,

It is estimated that as little as 1 to 3% of the bacte-

ria from a given environment can be cultivated �!. Tofully understand the complex world of microorganismsnew methods and technologies were needed.

This is where modern molecular biology enteredthe field of microbial ecology and aquatic biology.The 1990's saw the introduction and rapid advance-ment of new ways to examine organisms includingmicrobes. These methods included cloning, DNAsequencing, DNA fingerprinting and classifying organ-ism based on their DNA similarity rather than morpho-logical or physiological similarities,

An example of phylogenetic classification fornitrifying bacteria is presented in Fig 1. Ainmonia-oxidizing bacteria always begun with the prefixNitroso-! are confined to the beta subdivision of theProteobacreria, except for one bacterium, while nitrite-oxidizing bacteria begun with Nitro-! are more wide-spread with no members in the beta subdivision of theProreobacteria.

An initial goal of my research was to developmolecular probes for ¹trosomonas europaea andNitrobacter winogradsicyi. The probes would then beused to detect and quantify these bacteria from samples

taken from various filters and locations in functioningtest aquaria. Molecular probes are a short sequence ofDNA that can be made to match and so target! a spe-cific bacterium or a group or bacteria.

Once the probes were developed I would thenhave a novel way to quantify the nitrifying bacteria andcould begin the next phase of the project.

The second phase of the project was to use theprobes to determine what types of filter media did theAOB and NOB prefer, where do the AOB and NOBactually live on these inedia, and how close to eachother do the AOB and NOB reside. The answers to

these questions would allow one to develop better,more efficient biological filtration systems.

A NEGATIVE ANSWER

The results of my first 3 years of work were pub-lished in August of 1996 with a paper entitledComparative analysis of nitrifying bacteria associatedwith freshwater and marine aquaria co-authored byProfessor E.F. DeLong �!.

Briefly, I was able to develop a couple of molecu-lar probes for AOBs and one for NOBs, The AOBprobes were targeted towards AOBs such asNitrosomonas europaea and its closest relatives andNi trosospi ra AOB s. The NOB probe targetedNitrobacter winogradskyi and its closest relatives.

I could get the probes to work on pure cultures ofthe target AOBs or NOBs but sainples from freshwateraquaria always returned negative results. Initially, Ithought the problem was with the probes or the tech-nique used to extract the DNA or RNA from the aquar-iurn samples until I tested some samples from saltwateraquaria.

When I examined DNA from saltwater aquaria Igot a positive signal for the AOB probes but not theNOB probes. This meant that the extraction techniqueswere good. I then did an experiment the results ofwhich were presented at the 96th annual meeting of theAmerican Society for Microbiology held in NewOrleans, May 19-23, 1996, Fig. 2! where I set-up sixfreshwater aquaria and let then go through the estab-lishrnent of nitrification. When the ammonia added to

the tanks could be oxidized to nitrate in less than one

day I switched three of the tanks to saltwater andstarted three new saltwater tanks. I sampled andprobed all the tanks at different stages in the test.

The results of this test were that in no case could I

get a positive signal for freshwater aquaria from theprobes Fig. 2!. However, aH the saltwater aquaria,even the ones which had been freshwater, were positivewith my AOB probes while aH the freshwater aquariawere negative. Both types of aquaria were negative forthe Nitrobacter probe Fig. 2!,

These results meant that ammonia oxidation was

being done by different species of ammonia-oxidizingbacteria in fresh and saltwater aquaria. Otherwise theprobes would have detected AOBs in both environ-ments, Furthermore, while the AOBs in saltwater wererelated to Nitrosomonas europaea, the freshwater AOBwere most likely novel, Finally, nitrite-oxidation wasnot being performed by Nitrobacter winogradskyi or itsclose relatives in either environment but by someunknown nitrite-oxidizer,

While these results have upset the conventionalwisdom of people associated with the aquarium indus-try they were not shocking to microbial ecologists. Bythe time I had published my paper, a few other papershad been recently published regarding nitrifying bacte-ria.

Hiorns et al. �! in 1995 used molecular tech-niques to look for Nitrosomonas spp, and ¹trosospiraspp, and found that samples from soils and activatedsludge tested positive for Nitrosospira spp, but not¹trosomonas spp.. Further, when they examined lake-water and sediments they could not detect¹trosomonas spp. but did find Nitrosospira spp,These results suggested that the importance of¹trosomonas was over-emphasized. Obviously, otherAOBs were the primary ammonia-oxidizing bacteria inthese environments.

Wagner et al, �0! could not detect NitrobacterceHs in samples from river water, a trickle filter or acti-vated sludge and concluded that there were largenumbers or high-level activities! of non-¹trobacternitrite-oxidizing bacteria in the systems they examined.

Thus, there was plenty of other evidence, besidesmy work, at this time which pointed towards novelammonia- and nitrite-oxidizing bacteria being responsi-ble for nitrification in a wide range of aquatic environ-ments.

NITROSPIRA NOT NITROBA CTFR

My work continued and was now aiined at identi-fying the nitrite oxidizing bacteria in aquatic systems.The results were published in 1998 with a paper enti-tled ¹trospira-like bacteria associated with nitriteoxidation in freshwater aquaria with my co-authors L.T. Taylor, A. Blakis, and Professor E. F, DeLong �!,

When faced with the results that the bacteria ini-

tially thought to be present in a sample are not, one hasonly a few options on how to proceed. One option is toguess which bacteria might be present and designprobes for these bacteria, This is no better than trialand error. The second option initiaHy involves morework, but in the long run is really the only way toproceed, For this option, one develops a clone libraryfrom the sample. A clone library is a catalog of all thedifferent bacteria in a sample that is built by cloningand sequencing their DNA.

! Timothy A. Hovanec 2000 Page 3

F1G. 2. Results of the slot blotting and molecular probing for ammonia- and nitrite-oxidizing bacteria in freshwaterand saltwater aquaria. Presented at the 96th annual meeting of the American Society for Microbiology, NewOrleans, LA. The left side of the figure shows the mean N=3! aminonia top! and nitrite bottom! trendsfor freshwater aquaria, freshwater aquaria switched to seawater, and newly set-up seawater aquaria. Theright side shows the results of the slot blotting tests with moeclcular probes for all eubacteria eubac!, twoprobes for ammonia-oxidizers NlTROSO4E, NSM18! and nitrite-oxidizers NBAC2!. Only seawateraquaria show a positive signal for ammonia-oxidizers. No iVirrobacrer spp were detected.

! Timothy A. Hovanec 2000 Page 4

When you have an initial DNA sample from afilter or any other environment, the DNA consists of amixture of DNA from many different species. Theremay be the DNA from hundreds or more different bac-teria present in the sample. So the first objective is toseparate this inixed DNA sample into the individual

species DNA by cloning. Next you remove the DNAfrom the cloning vector, clean it up and sequence it.After you have developed sufficient sequence data, youcarefully check the sequence against the sequences ofknown bacteria using computer programs and a nationaldatabase, Finally, phylogenetic trees are constructed to

examine the relationship of the bacteria in your sampleto known bacteria see Figure 1!,

As is apparent the entire process is very time con-suming. Nevertheless, since this is the correctapproach it is the one I used to identify the nitrite-oxidizing bacteria in aquaria,

This work resulted in the discovery of new nitrite-oxidizing bacteria belonging to the phylum Nitrospira.In the clone library, I was only able to find NitrospiraNOBs, Conversely, I never found Nitrobacter NOBs,Once I had the correct NOB sequence, I developedunique inolecular probes for the Nitrospira NOB andprobed many freshwater and saltwater aquaria, Theresults showed that Nitrospira was present in relativelylarge numbers in all the aquarium samples butNitrobacter could not be found in any.

Finally, through a process called denaturing gradi-ent gel electrophoresis DGGE! I was able to track theappearance of the ¹trospira bacteria in newly set-upaquaria and quantify their growing numbers, relative toother bacteria in the sample, as the test aquarium cycled Fig. 5!.

CONFIRMATION BY OTHERS

The responses to the conclusions of my secondpaper varied depending on the respondent. Mostpeople associated with companies in the aquariumindustry did not and still do not! accept the results andpersist in their belief that Nitrosomonas europaea andNitrobacter winogradskyi are the nitrifiers in aquaria,Many of these companies sell preserved mixtures ofthese bacteria as aquarium starter cultures, It is import-ant to note, however, that no one associated with theindustry has published any contradictory data, peer-reviewed or not,

Among microbial ecologists and microbiologistswho are involved in this line of research the responsehas been much more positive. Furthermore, my con-clusions were strengthened by the publication in Mayof 1998 of yet another investigation of nitrite-oxidizingbacteria by Burrell et al. �!, These researchers, at theUniversity of Queensland, in Brisbane, Australia,investigated the microbial community of wastewatertreatment systems. Through the previously describedclone library development and sequencing techniquesthey found that bacteria belonging to the Nitrospiraphylum were the putative nitrite-oxidizing bacteria inwastewater systems.

The results of a third study on Nitrospira were alsopublished in 1998 8!. These researchers looked at themicrobiology of a nitrifying fluidized bed reactor usinganother molecular method called Fluorescent In Situ

Hybridization FISH!. Schramm et al. 8! could notfind any ainrnonia-oxidizing bacteria of the genusNitrosomonas or nitrite-oxidizing bacteria of the genus

Nitrobacter. Instead they determined that, in theirsamples, the nitrite-oxidizing bacteria were membersof the phylum Nitrospira and the ammonia-oxidizingbacteria were members of the genus ¹trosospi ra.

CONCLUSIONS

In the last few years, there have been a number ofpeer-reviewed papers published in leading scientificjournals by a wide array of international researchers oninvestigations of aquatic nitrifying bacteria from anumber of tested environments, There is a commonali-ty amongst these studies: namely, the lack of detectionof species of Nitrobacter and the finding of ¹trospiraas the nitrite-oxidizing bacteria, The simplest conclu-sion from this is that Nitrobacter are not, andNitrospira are, the nitrite-oxidizing bacteria in thesesituations.

In terms of ammonia-oxidizing bacteria, manynew species have been discovered and it seems likelythat the importance of Nitrosomonas europaea hasbeen over-emphasized, Meinbers of the genus¹trosospira and new members of genus Nitrosomonasare important ammonia-oxidizing bacteria in environ-ments ranging from ponds, to aquaculture facilities tofish aquaria.

The benefits of my research, and that of others, forthe aquatic filtration industry is that we can nowdevelop better biological filtration systems because wehave methods to detect and quantify the nitrifiers.Furthermore, the research calls into question the effica-cy and use of currently available starter cultures ofnitrifiers which are used to accelerate the establishment

of nitrification in aquaria, Many of these products arelabeled as containing Nitrosomonas europaea andNitrobacter winogradskyi while others do not name thebacteria species in their mixtures. The currentresearch of aquatic nitrifying bacteria would stronglysuggest that these products are of little to no benefit asthey do not contain the correct species of nitrifyingbacteria for closed aquatic systems.

REFERENCES

1. Amann, R. I., W. Ludwig, and K.-H. Schleifer. 1995.Phylogenetic identification and in situ detection of indi-vidual microbial cells without cultivation. Microbiol.Rev, S9;143-169.

2. Burrell, P. C., J. Keller, and L. L. Blackall. 1999.Microbiology of a nitrite-oxidizing bioreactor. Appl.Environ. Microbiol. 64:1878-1883.

3. Hlorns, W. D., R. Hastlng, C., I. M. Head, A. J.McCarthy, J. R. Saunders, R. W. Pickup, and G. H.Hall. 1995. Amplification of 16S ribosomal RNA genesof autotrophic ammonia-oxidizing bacteria demonstratesthe ubiquity of nitrosospiras in the environment.Microbiology. 141:2793-2800.

4. Hovanec, T. A., and E. F. DeLong. 1996. Comparative

@ Timothy A. Hovanec 2000 Page 5

Tlrrre sirree eqvarium start-up days!

Z k e 1 O

Tirade cleye!

FIG, 3. A! DGGE of select dates during the first 18 days after the startup of a freshwater aquarium, during which timenitrification became established. Clone 710-9, a iVitrospira-like putative NOB, can be seen to appear starting at about day 12 lane G!. B! Relative intensities of the band for clone 7! 0-9 at each sampling date. Froin Hovanec et al. 1998 �!,

H. Schleifer. 1996. In situ analysis of nitrifying bacteriain sewage treatment plants. Wat, Sci. Technol. 34:237-244.

analysis of nitrifying bacteria associated with freshwaterand marine aquaria. Appl. Environ. Microbiol. 62:2888-2896.

5. Hovanec, T, A., L. T. Taylor, A. Blakis, and E. F.DeLong. 1998. ¹trospira-like bacteria associated withnitrite oxidation in freshwater aquaria. Appl. Environ.M icrobiol. 64: 258-264.

Page 6! Timothy A. Hovanec 2000

5«s 5'

440

!nr «COCP

e

6. Kelly, D. P., S. Watson, and G. A. Zavarzin. 1989.Aerobic Chemolithotrophic bacteria and AssociatedOrganisms, p. 1807-1889. In J. T. Staley, M. P, Bryant,N. Pfennig, and J, G. Holt ed.!, Bergey's Manual ofSystematic Bacteriology, vol. 3. Williams 4 Wilkins,Baltimore.

7. Manz, W., R. Amann, W. Ludwig, M. Wagner, andK.-H. Scheifer. 1992. Phylogenetic oligodeoxynucleo-tide probes for the major subclasses of Proteobacteria:problems and solutions. Syst. Appl. Microbiol. 15:593-600,

8, Schramm, A., D. de Beer, M. Wagner, and R.Amann. 1998. Identification and activities in situ ofXitrosospira and ¹trospira spp. as dominant popula-tions in a nitrifying fluidized bed reactor. Appl. Environ.Microbial. 64:3480-3485,

9. Teske, A., E. Alm, J. Regan, S. Toze, B. Rittmann,and D. Stahl. 1994. Evolutionary relationships amongammonia- and nitrite-oxidizing bacteria, J. Bacteriol.176:6623-6630.

10, Wagner, M., G. Rath, H.-P. Koops, J. Flood, and K.-

11. Watson, S. W., and J. B. Waterbury. 1971.Characteristics of two marine nitrite oxidizing bacteria,¹trospina gracilis nov. gen. nov. sp. and ¹trococcustnobilis nov, gen. nov. sp. Arch, IVlicrobioI. 77:203-230,

12. Woese, C. R. 1987, Bacterial evolution. Microbiol. Rev.51:221-271.

Dr. Timothy A. Hovanec is the Chief ScienceOfficer for The Aquaria Group, the parent companyof Marineland Aquarium Products, AquariumSystems International and Perfecto Manufacturing.He is in charge of the biology, chemistry andmicrobiology laboratories at The Aquaria Group.

Voh 62, No. 8

Comparative Analysis of Nitrifying Bacteria Associatedwith Freshwater and Marine Aquaria

TIMOTHY A. HOVANECL * AND EDWARD F. DELONG'

Ecology, Evolution and Marine Biology Department, University of California, Santa Barbara,Santa Barbara, California 93I06, and Aquaria, Inc., Moorpark, California 9302I

Received 15 March 1996/Accepted 20 May 1996

Three nucleic acid probes, two for autotrophic ammonia-oxidizing bacteria of the P subdivision of the dassProteobacteria and one for Ix subdivision nitrite-oxidizing bacteria, were developed and used to study nitrifyingbacterial phylotypes associated with various treshwater and seawater aquarium biofilters. Xitrrrsornonas euro-paea and related species were detected in all nitrifying seawater systems and accounted for as much as 20% ofthe total eubactemd rRNA. In contrast, nitrifying bacteria belonging to the P-proteobacterial subdivision weredetected in only two satnples irom freshwater aquaria showing vigorous nitrification rates. rRNA originatingfrom nitrite-oxidizing ct subdivision proteobacteria was not detected in samples from either aquarium envi-rontnent. The data obtained indicate that chemolithotrophic ammonia oxidation in the freshwater aquaria wasnot due to p-proteobacterial phylotypes related to members of the genus Xurosornonas and their close relatives,the organisms usually implicated in freshwater nitrITication. It is likely that nitrification in natural environ-Inents is even more complex than nitrification in these simple systems and is less well characterized withregard to the microorganisms responsible.

2888

APPLIED AND ENvlRoNMENTAL MlcRQBIoLoGY, AIIg. 1996, p. 2888 � 28960099-2240/96/$04.00+0Copyright Ic> 1996, American Society for Microbiology

The pathways of the nitrogen cycle are highly dependent onmicrobial activities and transformations. One important path-way in the nitrogen cycle is nitrification, the oxidation of ain-monia to nitrite and subsequently to nitrate �7!, Traditionally,nitrification has been studied by chemical measurement ofammonia or nitrite disappearance, measurement of the pro-duction of nitrite or nitrate, ox a combination of these methods see reference 25 for a review of autotrophic nitrification!.Nitrification occurring in a wide range of environments, suchas soils �7!, ocean water �6!, freshwater lakes �1!, wastewa-ters �4!, and aquaria �6!, is assumed to be due to autotrophicbacteria. While heterotrophic nitrification can occur and maycontribute substantially to nitrification in certain environments�7, 29!, it is not coupled to energy generation and, therefore,is thought to be a minor component of overall nitrification �,25!.

A primary concern in fish culture systems ranging from high-density aquaculture operations to the home tropical fish aquar-iuin is the toxic effects of ammonia on fish, To control andmaintain safe ammonia levels in fish culture systems, biologicalfilters have been designed to promote the growth of ammonia-and nitrite-oxidizing bacteria. Biological filters usc a variety ofmaterials as supports on which the bacteria are cultured. Gen-erally, no special effort is made to distinguish between thetypes of supports used in diferent seawater or freshwater cul-ture systems. The general assumption is that species of ammo-nia- and nitrite-oxidizing bacteria are identical in the two typesof environments and that they require only a solid support,good aeration, and an energy source ammonia or nitrite! tobecome successfully established.

In freshwater systems, the bacterial genera responsible forthe oxidation of ammonia and nitrite are presumed to bepredominantly the genera Witrosomonas and lVitrobacter, bothof which are chemolithoautotrophic members of the class Pro-

' Corresponding author, Phone: 805! 529-1111. Fax: 805! 529-3030. Electronic mail address: hovaiiec@lifescLlscf,iicsb.ediI.

teobacteria �4, 38!. Recent studies in which comparative 16STRNA analyses of ammonia- and nitrite-oxidizing bacteriawere performed have clarified the phylogenetic relationshipsof these bacteria and have demonstrated that they belong totwo separate lineages within the Proteobacteria �2, 30!. Tcskeet al, �0! concluded that the nitrifying bacteria may havemultiple phylogenetic otigins. These authors speculated thatnitrifiers have developed independently many times, perhapsfrom diferent lineages of photosynthetic bacteria �0!. Thefreshwater autotrophic ammonia-oxidizing bacteria that havebeen characterized belong exclusively to the P subdivision ofthe Proteobacteria and are typified by Mtrosomonas eumpaea Fig, 1!. These bacteria form a distinct group within the J3subdivision and are affiliated with an iron~xidizing bacterium Gallionella fenuginea! and the photosynthetic bacteriuinRhodocyclus purpureus, along with methylotrophic bacteria.One ammonia oxidizer, Xitrosococcus oceanus, is a marinespecies that belongs to the y-proteobacterial lineage.

The most commonly studied autotrophic nitrite-oxidizingbacteria belong to the u subdivision of the Proteobacteri, ofwhich Xitrobacter Ivinogradskyi is a representative species Fig,1!. Other chemolithoautotrophic nitrite-oxidizing bacteria thathave been characterized are phylogenetically widespread in theclass Proteobacteria, occurring in the Ix, 8, and 7 subdivisions Fig. 1!. Phylogenetic analysis of the a subdivision of the Pro-teobacteria has shown that II/itrobacter Ivinogradskyi is mostclosely related to Bradyrhizobiumj aponicum and Rhodopseudo-monas palustris 9, 23, 27, 39!.

In this study, we used oligonucleotide probes which targetchemolithoautotrophic ammonia-oxidizing and nitrite-oxidiz-ing bacteria to examine nitrifying bacterial populations associ-ated with freshwater and marine aquaria. Various microbialhabitats associated with aquarium systems were investigated,including the gravel, water, and biofilter support medium,which is a substratum designed to encourage the growth ofnitrifying bacteria. Specific differences between nitrifying bac-terial assemblages on freshwater and seawater aquarium bio-filters were also investigated.

VoL. 62, 1996

NITROSO4Etar et rouI4BAC2 target group

NitrosomonasNitrosotobus multiformis

hll osovlbrlo tenulsIVltrososplrabrlensls

Nttrobam r Nitrobacter sp. LLNltrobecter hamburgesls

Rhodopseudomonas palustrist4Stfft 1 B target group

;: .=:�-'Plecacat'O'rfturM~<;,' Pseudomonas dimlnuta

Paracoccus deni tnll cansAfcatt'genes eutrophusComa monastestoslemni

hiyxococcvsxanthus'lvltiusococcus

oceanus

Pseureomonasaeruglnosa

'lVitrospinagracllls

Alteromonashaloplantttls hlltrcsplra

marina5.0tyya

Escherichia coll

Desutfovibre vulgaris

FIG. 1. Phylogenetic relationships of the chemolithoautotrophic ammonia- and nitrite-oxidizing bacteria. Most known aininonia-oxidizing autotrophs belong to theI3 subdivision of the Proteofxrcicria; the only exception is Nioosococcus oceanus, which is affiliated with the y subdivision. The nitrite-oxidizing bacteria are morewidespread in the Proteobectene, occurring in the a, 8, and y subdivisions. Nucleic acid probes which correspond to i! all known f3 subdivision ammonia oxidizers probeNITROSO4E!, ii! a clade on a deep branch in the f3 subdivision probe NSM1B!, and iii! the nitrite oxidizers belonging to the a subdivision probe NBAC2! weredeveloped. Nitrifying bacteria which are not targeted by the probes designed in this study are indicated by asterisks. Recent studies indicate that the genus Nioospiramay be affiliated with a group outside the 8 subdivision of the Proreohccterie, in a separate phylogenetic lineage 8!.

MATERIALS AND METHODS

Bacterial culture and nucleic add extrucuon techniques. Ammonia- and ni-trite-oxidizing bacteria were obtained from the American Type Culture Collec-tion or were kindly provided by J. B. Waterbury of Woods Hole OceanographicInstitute, Woods Hole, Mass., and were grown in organic-free media in batchculture by standard methods Table 1! �!.

Isolation of rt~ DNA genes of nitrite oxidizers. As expected, the nitrite-oxidizing bacteria grew slowly with low cell yields, and so the PCR was used togenerate sufficient ribosomal DNA template to test probe specificities. Prior tothe PCR, DNAs from Nioobacter winogradskyi and Nioobecrcr agilis were ex-tracted. Cells were placed in lysis buffer �0 mM EDTA, 50 mM Tris; pH 8.3! towhich lysozyme was added to a flnal concentration of 1 mg/rcl. After incubationat 37'C for 30 min, 50 p.l of proteinase K stock solution concentration, 10mg/ml! and 50 p,l of 20% sodium dodecyl sulfate SDS! were added to eachsample, and then the preparations were incubated at 55'C for 30 min. Cell lysiswas monitored by phase-contrast microscopy. Iu some cases, additional protein-ase K and SDS were added aud the sample was incubated at 55'C for another 30min.

After cell lysis, DNA was extracted by sequential extractions with phenol pH8.0!, phenol-chloroform-isoamyl alcohol �4:24:1!, and finally chloroform-isoamyl alcohol �4:1!. Each sample nucleic acid was precipitated with 0.3 Msodium acetate and 2 volumes of ethanol and stored at � 20'C. The sample wascollected by centrifugation, dried, and resuspended in 100 ttf of TE buffer �0mM Tris-HCI, 1 mM disodium EDTA!. The concentration of DNA was deter-mined by Hoechst type 33258 dye binding and fiuorometry model TKO 100minifluorometer; Hoefer Pharmacia Biotech Inc., San Francisco, Calif.!. Ribo-somal DNA was amplified by using primers specific for eubacterial rRNA, aspreviously described �!.

Isolation of rRNA. Cells of the ammonia-oxidizing and heterotrophic bacteriawere harvested by centrifugatiou for 20 miu at 8,000 rpm model RC5C centri-fuge; Sorvall Instrumenui!. Total rRNA was extracted 6'om bacterial cells by ceBdisruption with glass beads, using a Mini Beadbeater BioSpcc Products, Bartles-ville, Okla,!. After disruption, a three-step purification procedure with phenol Tris bufiered, pH 5,1], phenol-chloroform-isoamyl alcohol [24:24:I], and chlo-roform-isoamyl alcohol [24:I]! was performed �8!. The resulting crude nucleicacid was precipitated overnight at � 20'C after 0.1 volume of 3 M sodium acetateaud 2 volumes of ethanol were added. After precipitation, the nudeic acids were

NITRIFYING BACTERIA ON AQUARIUM BIOFILTERS 2889

collected by centrifugation and resuspended in 100 tsl of TE buffer pH 8.0!.RNA was quantified by measuring Arse with a Perkin-Elmer Lambda 3B spec-trophotometer by assuming that I A~ unit corresponds to 40 Isg of RNA per ml�8!.

Oligonucleotide probe design. 16S rRNA sequences of chemolithoautotrophicammoniamxidizing bacteria were aligned in a database by using sequence dataobtained from the Ribosomal Database Project �0!. Two regions were identifiedas having potential specificity for the target groups. One 20-nucleotide probe designated NITROSO4E! targeted all known ammonia-oxidizing mernbem ofthe fJ subdivision Fig. I!, and a second probe NSMIB! targeted three membersof the clade containing lVirrosomonas europaea, Nfrrosomones eutropha, andNitrosococ cur mob ilia

A third probe NBAC2! was designed to target the o subdivision nitrite-oxidizing bacteria Nioobecrcr w~kyi, Nttrobactcr egilis, and Wioobactcrhamixngensis. The probes were synthesized by Operon Tech, Inc., Alameda,Calif. The nucleotide sequences and positions of the probes are shown in Table2.

Probe hybridization procedures. To determine the specificity of each probe,probe binding to rRNAs from target and nontarget bacteria was monitored byautoradiography. A temperature series spanning the estimated dissociation tem-perature of each probe was used to determine the wash temperature empirically,

All probe hybridization experiments were conducted with a slot blot device Millipore Corp., New Bedford, Mass.!. rRNAs I'rom pure stock preparationsand samples were denatured with 3 volumes of 2% voflvol! glutaraldehyde andthen diluted to the final volume I:100! with dilution water � pg of polyriboad-enosine per liter, 0.0004% bromophenol blue!. The ptasmid stock preparationsof Nilrobccler wincgnrdrkyi and Ninobacrer agilis were diluted with an equalvolume of a mixture containing 1 N NaOH and 3 M NaCI. Samples were appliedto nylon filters Hybond N; Amersham Corp., Arlington Heights, Ill.! fitted intothe slot blot device. After air drying, the filters were cross-hnked by exposure to1,200 J of UV irradiation UV Stratalinker; Stratagene Corp., San Diego, Calif.!.

For hybridization experiments, membranes were placed in a heat-sealable bag,6 or 12 ml depcndiug on the number of membranes in the bag! of hybridizatioubufFer �,9 M NaCl, 50 mM Nape, 5 mM EDTA, 0.5% SDS, 10X Denhardt'ssolution, 0.5 mg of polyadeuosiue per mi! was added, and the bag was sealed audplaced in a hybridization oven model 136500; Boekel Industries, Inc,! for 30 miuat 45'C, After 30 min, the bags were removed, and 2 x 10 cpm of P-end-

2890 HOVANEC AND DLLONG APPI- ENvlRoN. MICROBIoL.

TABLE 1. Sources of the bacteria utilized in the nucleic acid probe validation studies and culture media used to grow themProteobacte rial

subdivision Strain Growth mediumSpecies

ATCC 19718NC2 Waterbury'ATCC 25196C128 Waterbury'NV1 2 Waterbury'ATCC 19707

ATCC 221ATCC 928 �5%!ATCC 929ATCC 221ATCC 929ATCC 928

BetaBetaBetaBetaBetaGamma

AlphaAlphaGammaDeltaDelta

ATCC 25391ATCC 14123ATCC 25380NB295 WaterburyNB211 Waterbury"

ATCC 480ATCC 96ATCC 481ATCC480ATCC 480

Luria-BertaniLuria-BertaniLuria-BertaniLuria-BertaniLuria-BertaniLuria-BertaniLuria-BertaniLuria-BertaniMarine brothLuria-Bertani

BetaBetaBetaBetaAlphaAlphaAlphaGammaGammaGamma

ATCCATCCATCCATCCATCCATCC501'ATCCATCCATCC

17697155541566811975'1774117001

80712713217503

' Kindly provided by J. B. Waterbury, Woods Hole Oceanographic institute.s The rnedimn used was 25% ATCC 928 medium in distilled water.' Received from P. Baumann.

TABLE 2. Nucleotide sequences and positions of the three oligonucleotide probes for nitrifying bacteria

T *C!'/ Nontarget bacteria withPost tlon Base sequence�1 3/! Targeted group exact match to probe

sequence

l3-Proteobacteriaiammonia oxidizers

Nilrosom onus europaea,Ni lrosomonaseutnrpha,¹rosococcus mobilis

Nilrobacter winognzdskyi,Nilrobacterhamburgensis,Nitrobacler agilis

Nodularia sp.'

Nones

NITROSO4E 639-658 CAC TCT AGC YTT GTA GTT TC 43.2/53.0

41.2/53.0479-495 TCT GTC GGT ACC GTC ATNSM1 B

Afipia clevelandensis,Afipia fehs,Rhodopseudomonaspaluslns strain,Bra dyrhizobium

j aponi crun'

1017-1036.1 GCT CCG AAG AGA AGG TCA CA 49.4/53.0

Zscheriehia cali numbering.T�, dissociation temperature. Wash temp, experimentall determined wash temperature see Materials and Methods!.

' The following two nontarget bacteria have a one-hase mismatch with the probe sequence: Oscillarona sp. and Cytindrospcrmum sp."There are 50 nontarget bacteria or strains of bacteria that have a one-base mismatch with the probe sequence, These bacteria include Ehrltchfa, tfhadovutum,

Rhodobacrer, Rhodoptanes, and Fusobactenum species, as wefi as Anaptarma marginate, 1Mobaciltus Ihioparur, Sebaldella rermiridis, and Srreprobacillus monitifonnir.The fofiowins tour nontarget bacteria have a one-base mismatch with the probe sequence: Photorhizobium rhompsonianum, Phororhizabium sp. strain IRBG 230,

fi~ium sp., snd Phorortdzobaun sp. strain MKAa 2.

Chemolithoautotrophic ammonia-oxidizing bacteriaNitrosomonas europaeaNitrosococcus mobilisNi trosolobus mrduformisNilrosospira bri ensisNi lrosovibno tenuisNi lrosococcus oceanus

Chemolithoautotrophic nitrite-oxidizing bacteriaNitrobacler winogradskyiNitrobacler agilisNilrococcus mobdisNfrrospiru marina¹rospina gracilis

Heterotrophic bacteria closely related to ammonia or nitrite oxidizeersAlcaligenes eulrophusA lcaligenes faecaiisComamonas acidovoransComamonas teslosteroniPamcoccus denitrificansRhodopseudomonas paluslnsPseudomonas diminutaShewanelta pulrefaciensPseudomonas na uticaPseudomonas aeruginosa

labeled probe was added. Each bag wss resealed and returned to the oven. Themembranes were incubated overnight in the hybridization oven at 45'C.

After the overnight washing described above, the membranes were removedand washed in a solution containing 1x SET l50 mM NaCl, 1 mM EDTA, 20mM Tris; pH 7.8! and 1% SDS at room temperature for 30 min on a shaker table.The membranes were then washed in fresh Ix SET � 1% SDS at appropriatewash temperatures for 30 miu with shaking every 10 min!. After washing, themembranes were allowed to air dry. Autoradiographic signals were quantified byusing a gss proportional radioisotope detection system Ambis, Ine., San Diego,Calif.!. Film autorsdiographs were aho recorded with an intensifier screen for 20to 24 h at � 76'C.

The relative rRNA-speeiiie hybridization signal attributable to each probe wasdetermined by calculating a slope couuts per minute bound per nanograrn ofRNA! for the serially diluted sample. Values were normalized by using a cor-

rection factor determined by dividing the group-specific probe slope derivedfrom known rRNA standards by the slope derived from the eubacterial probe forthe same standards �0!. Group-specific hybridization signal was calculated bydividing the normalized group-specific probe slope by the eubacterial probeslope of the same sample.

Sampling and extraction of nueteic acids tram aquarium samples. A variety oflocations in small water volume, �00 liters! aquaria having two general typesof environments inorganic and organic! were sampled for the presence ofchemolithoautotrophic uitrifying bacteria.

The samples consisted of aquarium gravel, aquarium water, and pieces of theaquarium biological filter media Gravel wzs collected with a scoop, weighed tothe nearest 0.1 g, placed in a polypropylene tube, and immediately covered withlow-pH buffer �0 mM sodium acetate, 10 rnlvl disodium EDTA! for rRNAextraction or with cell lysis buifer for DNA extraction. Samples were stored at

VoE 62, 1996

b c

jt iIthl,ll,' '; ','...

6 ~ 8! N1TROSO4E A! EU BACTERIA

C! NSM18 �! NBACZ

RESULTS

� 20'C until extraction, Aquarium water was collected in prewashed glass jarsand filtered through a Sterivex GV filter by using autoclaved pump tubing and aperistaltic pump. Between 1,000 and 4,000 ml of water was filtere depending onthe sample. After filtering, l.g ml of cell lysis butfer was added to each unit witha sterile syringe, and the filters were stored at � 20'C until processing, Variousbiological filter media were collected by cutting a piece of material I'rom the filterwith alcohol-sterilized scissors and forceps. Each medium sample was placed ina polypropylene tube, covered with 2.0 to 2.5 ml of cell lysis bufi'er or beadbeating solution, and then stored at � 21FC until extraction.

The gravel samples were extracted by adding 200 Icl of 20% SDS and 3 ml ofphenol Tris bufi'cred to pH 5.1! and shaking the preparations by hand for 5 min;this was followed by incubation in a 60'C water bath for 7 min. After shaking for3 min, the samples werc centrifuged at 1,500 ipm for 10 min model 1550centrifuge; Hamilton Bell, Montvale, Nik!. The nucleic acids were aliquoted intothree tubes, and the contents of each tube were extracted by using the beadbeating protocol described above,

The nucleic acids in the aquarium water samples were extracted by adding 40Icl of lysozyme from a stock solution containing 25 mg of lysozyme in 500 Id ofdistilled water! to each thawed sample, The filter was placed on an agitator andshaken at 37'C for 30 min. Then 500 p,l of proteinase K stock solution concen-tration, 10 mg/ml! was added, and the filter was incubated at 55 C for I h withshaking. The solution was drawn out of the Sterivex fdter with a syringe into apolypropylene tube. Phenol-chloroform-isoamyl alcohol extraction was per-formed, and this was followed by a series of chloroform-isoamyl alcohol cxtrac-tions. The solution was concentrated with a Centricon 100 concentrator Ami-con, Beverly, Mass.!, and nucleic acids were precipitated.

Freshwater and seawater aquarium blofilter comparison. Six all-glass aquaria capacity, 34 liters! were used along with a standard home aquarium filtrationsystem Penguin model 160B; Marineland Aquarium Products, Moorpark, Ca-lif.!. There was no substratum or other material in the aquaria. In the model160B system the main body of the filter unit hangs on the outside upper backedge of the aquarium. On the upper weir of the filter unit is the dedicatedbiological filter BioWheel; referred to below as the biofilter!, which sits per-pendicular to the water iiowing back into the aquarium. The water fiow causesthe biofilter to continuously rotate such that it functions as a rotating biologicalcontactor, and, therefore, the filter surface alternates between a partially sub-merged phase and an air-exposed phase.

Initially, the tanks were filled with dechlorinated activated carbon-treated!tap water; 5 mM ammonia made with ammonium chloride! was added to eachaquarium daily for the first 20 days and then every other day or so. Aquariumwater was sampled several times a week and was analyzed by performing a fiowinjection analysis FIAstar system; Tecator AB, Hoganas, Sweden! for ammonia gas diffusion membrane method!, nitrite azo dye method!, nitrate cadiniumreduction-azo dye method!, and acid-neutralizing capacity methyl orange to anend point of pH 4.5! as recommended in the manufacturer's application notes,The pH was determined with an electrode and a specific ion meter OrionInstruments!.

After all of the aquaria were exhibiting nitrification, as determined by nitrateproduction, the water in one group of three aquaria was changed from frcshwatcrto seawater prepared with artificial sea salts [Marineland Commercial Aquari-ums, Moorpark, Calif.j!. Three additional aquaria were also set up with artificialseawater and filter units with BioWheels which had never been run. Waterquality data were collected for the nine aquaria as previously described foranother 75 days. At 43 and 72 days after the one freshwater group had beenswitched to seawater, the biofilm on each biofilter was sampled by cutting out asmall piece of the filter. rRNA and ribosomal DNA were extracted as describedabove. rRNA was analyzed by using oligonudeotide probes as described above.

Oligottucleotide probe specificity. The specificities of threeof the four ammonia- or nitrite-oxidizing group-specific probesdeveloped IB this study are shown IB Fig. 2, Database searchesarid hybridization experiments performed with rRNAs ex-tracted from phylogeitetically diverse bacteria indicated thatthe probes were sUIIIcierttly specific to identify various chemo-lithoautotrophic nitrifying bacteria with the following pfovi-sioits. There is one norttarget organism for the IvtITROs04Eprobe and there are four itoittarget organisms for the NBAC2probe jB which the probe sequence compliments the Itontargetsequence exactly Table 2!. III the case of probe NSMIB thesequences of about 50 itorttarget organisriis Ottt of the databaseof more than 3,000 sequences have only one misiitatch with thetarget sequence Table 2!.

A range of wash temperatures was tested to determine theoptimal conditions for probe specificity. Under appropriatehybridization aitd WaSh COnditions, the XITROSO4E PrObe

NITRIFYING BACTERIA ON AQUARIUM BIOFILTERS 2891

a b c 3 b

5

!j;','.l:::::: i:-,:,':-: 1:;,,::.:-'::.":::.: .

7

FIG. 2. Autoradiographs demonstrating the specificity of eubacterial probeEUBAC A!, ammonia-oxidizing bacterial probes NITROSO4E B! andNSM1B C!, and nitrite-oxidizing bacterial probe NBAC2 D!. The rRNA ex-tracts from chemolithoautotrophic nitrifying bacteria and closely related bacteriawere blotted in the slots in the following arrangement: slot a-l, Niriobacrerrrinogradskyi; slot a-2, Nirrobacier agilis; slot a-3, Nirrosovibrio renuis; slot a-4,Nisrosospba bnensis; slot a-5, Nirrosolobus mulrfonnis; slot a-6, Niirosomonaseuropaea; slot a-7, Nirrosococcus mobilis; slot b-1, Rftodopseudomonas palusrrfs;slot b-2, Pseudomonas diminuta; slot b-3, Paracoccus denirnficans; slot b-4, Co-mamonas acidovarans; slot b-5, Alcafigenes faecafis; slot b-6, Comamonas resros-reroni; slot b-7, Afcaligenes euirophus; slot c-l, Nirrococcus mobilir; slot c-2,Niirosococcus oceanus; slot c-3, Shewaneffa purrefaciens; slot c-4, Pseudomonasnaudca; slot c-5, Pseudomonas aeruginosa; slot c-6, Nirrospina gracilis; slot c-7,Nllrosplra rriarina.

bound the rRKAs of a!l of the P subdiviSiOn ammoitia-oxidiz-ittg bacteria examined, but ttot the rRNAs of the closely relatedheterotrophic bacteria Fig. 2!. The NSMIB probe yieldedpOSitive signals with the two targeted P subdivision ammonia-oxidizing bacteria Nitrosomonas europaea aitd Nifrosococcusmobilis! but itot with other nitrifying bacteria or closely relatedheterotrophic bacteria belonging to the same subdivision Fig.2!. We tested a third probe for the P subdivision oxidizers,XLBI, bitt this probe cross-reacted with the closely relatedheterotrophic bacteria at all wash temperatures tested dataitot shown!. None of the nucleic acid probes designed for theP subdivision ammonia-oxidizing bacteria hybridized to ¹-frvsococcus oceanus, a marine species which is the only hlownautotrophic ammonia oxidizer riot IB the I5 subdivision Fig. 2!.

The results of the specificity test for the nitrite-oxidizingbacterial probe NBAC2! show that this probe is specific forIWO knOWn tx SubdiViSiOn nitrite OXidiZerS NitrObrSCter WinO-gradskyi and Xifrobacter agilis! aitd does riot cross-hybridizewith either the 5 or the y subdivision nitrite-oxidizing bacteria

2892 HOVANEC AND DELONG APPL. ENVIRON. MICROBIOL.

or closely related n subdivision heterotrophic bacteria, such asRhodopseudomonas paiusrris Fig. 2!.

The data indicated that three of the four nucleic acid probestested were suificiently specific to distinguish autotrophic ni-trifying bacteria from closely related heterotrophic bacterialspecies.

Detection of autotrophic nitrifying bacteria in aquaria. Wetested the nitrifying bacterial rRNA probes with nucleic acidsextracted from a wide range of samples obtained from activelynitrifying freshwater and seawater aquaria {Table 3!. Some ofthe samples came from biofiiters in aquaria which receivedmore than 82 g of fish food or were dosed with 32 mM am-monia each day. Only 2 of the 38 freshwater sainples gave apositive result with any of the nitrifier-specific probes Table3!. These two samples, which exhibited positive signals for thetwo ammonia oxidizer rRNA probes, were from biofilterswhich had been dosed with ammonium chloride and werenever exposed to the fish waste or organic compounds that arenormally associated with a fish tank. The positive signals ob-tained for these biofilters may have resulted from contamina-tion from seawater biofilters located nearby. These biofiltershad been in the culturing system for 76 days before sampling.The NBAC2 probe did not indicate the presence of 0I subdi-vision nitrite-oxidizing bacteria in any of the freshwater sam-ples Table 3!, There were large amounts of eubacterial rRNAdetected by the eubacterial probe in each sample, and so thelack of signal cannot be attributed to insuBicient material onthe membranes.

A PCR analysis in which two general eubacterial primers forward primer 8-27 and reverse primer 1492-1510! �9! wereused was performed with some samples to increase sensitivityand to determine whether nitrifying bacterium rRNA genescould be detected in the mixed-community DNA. PCR prod-ucts were blotted, and hybridization experiments were per-formed with the nitrifier rRNA probes. No signal was detectedin the PCR products, which is consistent with the results of therRNA hybridization experiments.

Positive results with probes specific for ammonia-oxidizingnitrifiers NITROSO4E and NSM1B! were obtained for allseawater samples, which were dosed daily with ammoniumchloride Table 3!. The lengths of time in the systems for theseawater biofilters tested ranged from 53 to 299 days. As withthe freshwater systems, negative results were obtained with theprobe for nitrite-oxidizing bacteria NBAC2!. Quantitative oli-gonucleotide probe hybridization experiments indicated thatas much as 20% of the eubacterial rRNA was derived fromammonia-oxidizing bacteria belonging to the P subdivision ofthe Proteobacteria Table 4!. This is consistent with the pre-sumed presence of significant numbers of ammonia-oxidizingbacteria on the biofilters. Furthermore, since the signal of theNI'tmsomonas species probe NSMl B! is equivalent to the sig-nal of the more general P-proteobacteriai ammonia-oxidizinggroup-specific probe NITROSO4E!, the nitrifiers on the sea-water biofilters appear to be dominated by NI'Irosomonas eu-ropaea and its close relatives rather than ¹rrosospira types.

Freshwater-seawater biofilter comparison. The mean am-monia, nitrite, and pH data for the three groups of biofiltersfrom aquaria that received different water treatments are pre-sented in Fig, 3, It is clear that established freshwater aquar-ium biofllters experienced a complete loss of nitrification whenthe water in the aquaria was switched to seawater. This causedan increase in the ammonia concentrations in the aquaria Fig.3!. After the switch to seawater, it took the previously Aesh-water biofilters nearly as long to reestablish ammonia oxida-tion as it took the newly set up seaivater biofilters. However,the maximum ammonia concentration reached during the es-

tablishment period was less in the switched biofilters than inthe newly set up seawater biofilters Fig, 3!, There was a small,temporary increase in the ammonia concentration in the fresh-water aquaria from day 9 to day 17 after the switch!, whichcoincided with a drop in the pH to less than 7.00. The pH rose and ammonia disappeared! after the addition of NaHCOs.~ Nitrite oxidation was established faster in the newly set upseawater biofilters than in the biofilters switched from fresh-water, with complete oxidation occurring by day 50 and by day60 {after the switch!, respectively Fig. 3!. Furthermore, thenitrite concentration reached a much higher value and re-mained higher for a longer period of time in the switchedbiofilters than in the newly set up seawater biofilters. Thenitrite concentration in the continuously freshwater biofilterswas low for the duration of the measuring period Fig. 3!. Apartial water change was performed on day 29 after theswitch! in all aquaria, and this change is refiected by the sud-den drop in the nitrite concentrations in the seawater andfreshwater-to-seawater groups. The nitrite concentrationsteadily increased again after day 29 in both groups until itfinally decreased before the end of the measuring period be-cause of establishment of nitrite oxidation.

The pH trends for the three groups of biofilters were similarexcept for a period of 8 days early in the test days 9 to 17!when the pH in the freshwater biofilter group fell to less than7.00. This pH change was compensated for by the addition ofNaHCO3.

Oligonucleotide probe hybridization experiments revealedpositive signals with both ammonia-oxidizing bacterial probesfor all seawater filter regardless of age newly set up filters andfilters switched from freshwater! Fig, 4!, Freshwater biofiltersconsistently yielded negative results with all of the nitrifier-specific probes Fig. 4!. The results indicated that NI'rrosomo-nas europaea or its close relatives were well represented on theseawater biofilters. The results obtained with the probe fornitrite-oxidizing bacteria were negative for all samples Fig, 4!.Thus, both Nirrobacter Ivinogradskyi and Nirrobacrer agilis wereeither absent or present only at concentrations below our limitsof detection, even though the nitrate concentrations steadilyincreased during the test.

DISCUSSION

Definitive studies correlating nitrification rates with nitrify-ing microorganisms in natural samples are difficult, Until re-cently there were few available methods for identifying andquantifying specific bacteria or groups of bacteria in environ-mental samples without cultivation, an approach known tosometimes lead to biased representation �, 32!. Cultivation ofnitrifying bacteria is especially challenging because of the slowgrowth rates of these bacteria and the frequent occurrence ofculture contamination by heterotrophic bacteria �2, 31!, Ward�5! utilized immunofluorescence techniques to enmnerate ni-trifying bacteria, but this technique also required cultivation ofthe target group to raise antibodies. More recently, PCR prim-ers have been developed and used to detect ¹rosomonas spp.,Nitrosospira spp., and ¹trobacter spp. in diverse environments�, 13, 21, 22, 31!. Wagner et al, �3! developed fiuorescent insitu hybridization probes specific for certain I3-subdivision pro-teobacterial ainmonia oxidizers. These authors found that upto 20% of the total bacteria in activated sludge samples froman animal waste-processing facility could be ammonia oxidizers.

In this study, oligonucleotide probes were used successfidlyto detect ammonia-oxidizing chemolithoautotrophic bacteriain environmental samples i,e�seawater aquarium biofilters!.Furthermore, the data obtained indicated that the bacteriaresponsible for ammonia oxidation in fieshwater aquaria are

NITRIFYING BACTERIA ON AQUARIUM BIOFILTERS 2893Vot., 62, 1996

TABLE 3. Results of probing rRNAs extracted from biofilms attached to various aquarium biofiltration media or aquarium water withdomain- and group-specific oligonucleotide probes

Signal detected by the following oligonucleotide probes r:

Bacterial NITROSO4E NSMIB NBAC2Ammoniasource

Bioffimsubstrate

Aquariumenvironment

Daily amt ofamrnoniaaSample

+ + +

+ + + + ++ + + + +

++

+ + + ++ + +

+ + + + +' A continued ne@gibte concentration of ammonia in the rystenu which bad daily inputs of fish food or ammonium chloride was considered evidence that nitrification

occurred.a The type of aquarium water.' The medium irom which the bacterial cells were extracted,

The values in grams are the amounts of fish food put into the aquaria each day; the molar and millimolar valucx indicate the amounts of ammonia added to theaquaria or systems in which the biofilters were located each day.

' Fish means that the aquarium had a fish population and ammonia was generated by the fish; NH~C1 means that there were no fish in the tank and the ammoniawas 1'rom ammonium chloride added daily.

r +, signal detected; �, no signal detected.

which is consistent with the results of previous studies. How-ever, previously characterized t3 subdivision aminonia-oxidiz-ing bacteria were detected in vigorously nitrifying freshwateraquaria in only 2 of 38 samples.

different from the bacteria responsible for ammonia oxidationin seawater aquaria, In seawater aquaria, Xifrosornortas euro-paea and related phylotypes appear to be present at high levelsand are presumably the active ammonia-oxidizing bacteria,

13011302130313041306130713091312131513167501750275037504710r711rCAQBWCAQBWEST32BEST33BEST34BFlwrte5FlwrtegFlwrte9FWSW4FWSW6MejBW-AMejBW-BT408T408T825T825WDF1025WDF1 026WDF1036WDF1036WDF1039WDF1039714r715rFWSW2FWSW3FWSWSFWSW9SW117SW123SW129SW134SW148SW152SW159SW202

FreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawater

Bulk waterGravelGravelFilter fiberBulk waterGravelPolypropyleneBulk waterBulk waterPolyfiberPolyfiberBulk waterPolyfiberPolyfiberPolyfiberPolyiiberPolyfiberPolyfiberSpongePolyprapyleneFilter fiberGravelGravelGravelPolyprapylenePolypropylenePolyprapylenePolypropyleneDetritusGravelGravelGravelSpongePolypropylenePolypropyleneGravelPolyp ropyleneGravelPolyfiberPolyfiberPalypropylenePolypropylenePolypropylenePalypropylenePolyfiberPolyfiberPolyfiberPolyfiberPolyfiberPolyfiberPolyfiberPolyfiber

4g4g4g4g4g4g4g32.1 mM32.1 mM32.1 mM32.1 mM82.84 g32.1 mM82.84 g32.1 mM32.1 mM32,1 inM32.1 inM1,4 g1.4 g1.4 g10 mM10 mM10 mM5 mM5mM82.8 g82,8 g3.5 g3,5 g0.8 g0.8 g2.0 g2.0 g3.2 g3,2 g3.2 g3.2 g714 mM714 mM5 mM5 mM5 mM5 mM2,5 mol2.5 mol2.5 mol2.5 mol2.5 mol2.5 mol2.5 mol2.5 mol

FishFishFishFishFishFishFishNH,C1NHxClNH~CINH4ClFish~CIFishNHxClNH4ClNH4C1NH4C1FishFishFishNH4ClNH4ClNH4ClNH4ClNH ClFishFishFishFishFishFishFishFishFishFishFishFishNHxClNHxClNH,ClNH4CINH,ClNH,ClNH4CINH,CINHxClNH,CIXHxCINHxC1NHxC1NHxC1

+ + + + + ++

+ + + + + ++ ++ ++ + + + + + + + + + + + + + + + + + + + + + + + + + + + +

APPL. ENvIRON. MIcROBIOL.

a b c b c'I i','i'>-",",' *, ',tt*c l!,'"*'fi

3': i'.:::--":::,-:-::i:�'::::::::,:::.f '.'"'

zz;" .,-".t'

B! N TROSO4E

freshwater «hahgeato seawater hlott1ter

% Hybridization to: freshwater hier herAgc of biofilter Daily ammonia

d ys! dose mol! NITROSO4E NSM1Bprobe probe seawater bio ster

SW202 538W148 98

2.52.5

20.4 23.818.5 17.4

A! EUBACTER A

e b c

C! h S �1 B D! NBt C21.00.90.80.7

E 0.6

cI 0.4E

0.30.20.'I0.0

2.50E

2.00

1.50

0AO

2894 HOVANEC AND DELONG

TABLE 4. Lcvcls of hybridization normalized to the cubacterielprobe! of the two probes specific for �-subdivision

ammonia-oxidizing bacteria to rRNAs extracted fromthe biofilms of two seawater aquarium filters'

' The biofilters were part of a larger group of 35 filters dosed daily with 2.5 molof ammonia, The NITROSO4E probe targets afi � ammonia oxidizcrs, while theNSM1B probe targets a sub croup of these bacteria Fig, 1!.

There are three possible explanations for our observations; i! there were few nitrifiers relative to other bacteria in thesamples examined, and so the method used was not sensitiveenough; ii! heterotrophic bacteria were responsible for theoxidation of ammonia and nitrite in the environments studied;

8.007.507.006.50

6.00 -15-10 -5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75Time days!

FIG. 3. Mean values n = 3! for ammonia concentration A!, nitrite concen-tration B!, and pH C! for the biofiltcrs from the following three aquariumenvironments: 0'cshwatcr «hanged to seawater f3!, freshwater C !, and seawater L!. Bars indicate standard errors, Each aquarium received 5 mM ammonia asammonium chloride! each day for the first 20 days and then nearly every otherday; none of the aqua« a contained fish. Establishment of nitrification is shown bythe sudden doc«case in the ammonia concentration for the seawater group andthe group in which I'rcshwatcr was changed to seawater near day 15 aftcr theswitch!. This was followed by a rapid decrease in the nitrite concentrationbetween days 45 and 60. A partial water chango wss performed on all aquaria onday 29, and this resulted in the large temporary decrease in nitrite concentrationevident at this time.

a b,, c

freshwater «hehgec

3:--::.:::::-::-:-:::::::.':::::.':::::::::. � 3freShWater htetrlter4 ' i':.':..:'.: ': ''-': 4

seewete h ottker 5

FIG. 4. Slot blot analysis of rRNAs extracted from the biofilters of twofreshwater aquaria changed to seawater rows 1 and 2!, two continuously I'rcsh-water aquatia rows 3 and 4!, and two seawater aquaria rows 5 and 6! andhybridized with the eubacterlal probe A!, the NITROSO4E probe B!, theNSMIB probe C!, and the NBAC2 probe D!. Water chemistry was tested threetimes a week for these filtcr units scc Fig. 3!, and the data confirme that activenitrification occurred. Lanes a, «RNA samples taken 43 days after the switchfrom fresh water to seawater; lanes b, rRNA samples taken 72 days after theswitch; lanes c, rRNAs extracted 0'om control strains slot c-l, Nt'trosomonaseuropaea; slot c-2, Comarnonas testosteroni; slot c-3, Nitroba«ter winogradskyi; slot«-4, Rhodopseudomonas patastrts!.

or iii! the responsible species of autotrophic ammonia-oxi-dizing bacteria belong to another phylogenetic group whichthe probes did not detect. These possibilities are discussedbelow.

The minimum detection limit for radiolabelled nucleic acidprobes is between approximately 0.1 and 1.0% of the totalrRNA �!. While absolute bacterial cell numbers cannot beinferred from the results of hybridization experiments, thismethod does provide a reasonable indication of the relativebiomass or metabolic activity of the targeted group, The bio-filter experiments demonstrated that our method was suffi-ciently sensitive to detect nitrifiers in this environtnent, sinceall seawater samples produced a strong positive signal. It isreasonable to assume that in the aquaria, whose sole energyinput was ainmonia, the bacteria responsible for nitrificationwere active and constituted a large fraction of the total bacte-rial assemblage, The positive results obtained with the autotro-phic ammonia-oxidizing bacterial probes for seawater biofiltersexposed to the same environmental conditions as parallelfreshwater filters indicate that the extraction and hybridizationprocedures were sufficiently sensitive to detect ammonia oxi-dizers belonging to the P-proteobacteriai subdivision on filters.

Heterotrophic nitrification has been shown to be potentiallygreater than autotrophic nitrification in certain environments�6!. There are several species of heterotrophic bacteria whichuse ammonia as a substrate and produce either nitrite, nitrate,or a less common nitrogen cycle intermediate, such as hydrox-ylamine lg!, Tate �9! found insufficient numbers of ¹tro-somonas and ¹trobacter cells to account for the nitrate pro-duction in histosols. Instead, using inhibitors, he determinedthat an Arthrobacter population was responsible for a major

VOL. 62, 1996 NITRIFYING BACTERIA ON AQUARIUM BIOFILTERS 2895

portion of the nitrification occurring in these soils. Castignettiand Hollocher �! identified six heterotrophic bacteria, includ-ing Pseudomonas denitrificans, Pseudomonas aeruginosa, andtwo strains of Pseudomonas fhorescens, that exhibited nitrifi-cation activity. While there is no direct evidence that hetero-trophic nitrification was the dominant process in the presentstudy, this possibility cannot be totally discounted. However,such heterotrophic nitrification seems unlikely since theaquaria received only inorganic ammonia ammonium chlo-ride! as an energy source. Carbon dioxide was the sole carbonsource available as there were no significant inputs of organiccarbon to support heterotrophic bacterial growth beyond tracecontamination. It is doubtful that heterotrophic bacterialgrowth was significant in this lithotrophic environment.

The possibility that the autotrophic nitrifying bacteria in thefreshwater aquaria studied belong to subdivisions other thanthe P subdivision of the Proteobacteria seems to be the mostlikely explanation for our observations. The probes were sen-sitive enough to detect ammonia-oxidizing bacteria in seawatersystems. The freshwater systems had similar rates of nitrifica-tion, but ammonia oxidizers were not detected. Furthermore,iVi trosomonas spp. were detected on biofilters only after a shiftfrom freshwater to seawater. This suggests that there werechanges in the population of nitrifying bacteria, as indicated bythe appearance of ¹rosomonas spp. and their relatives, aswell as the transient decrease in nitrification immediately fol-lowing the shift from freshwater to seawater. In total, our datasuggest that microorganisms other than the usually implicatednitrifiers members of the l3 subdivision of the Proteobacteria,such as ¹trosotnonas spp. and their relatives! are the majoragents responsible for nitrification in the freshwater aquariumenvironments examined,

To date, only one ammonia-oxidizing bacterium which doesnot belong to the P subdivision of the Proteobacteria has beencultured, The emphasis on ¹trosomonas types, especially ¹-trosomonas europaea, as the major ammonia oxidizers in envi-ronments may be partially a result of culture bias. It is possiblethat iVitrosomonas europaea grows better in enrichment cul-tures and pure cultures than other, more ecologically signifi-cant nitrifiers which fiourish and outcompete ¹trosomonasspp. in mixed populations, Belser and Schmidt �! observedselectivity among the dHferent genera of ammonia oxidizers,with Nitrosomonas spp, generally dominant over iVitrosospiraand Witrosolobus spp., possibly because of a faster growth rate.Furthermore, these authors found that while a medium couldsupport the growth of either ¹trosomonas species or Xi-trosospira species, these bacteria generally never grew togetherin the same enrichment culture. This may explain the data ofHiorns et al, �3!, who suggested that ¹trosomonas spp. wereprevalent only in enrichment cultures that were not obtainedfrom environmental samples.

In the case of nitrite-oxidizing bacteria, the data suggest thepossibility that the responsible bacteria were not ¹trobacterspecies, since we were unable to detect Nitrobacter cells in anysample examined, Nitrite-oxidizing a subdivision proteobacte-ria were also not detected by Wagner et al. �4!, who examinedriver water, a nitrifying trickle filter biofihn, and activatedsludge samples by using fluorescent probes specific for variousWitrobacter species. These authors concluded that the mostprobable reason for their results was that there were largenumbers or high-level activities! of non-¹trobacter nitrite-oxidizing bacteria present in the systems which they examined,The chance that our results were due to the relatively low levelof sensitivity of quantitative rRNA hybridization experimentsdoes exist. However, DNAs from aquarium samples amplifiedby PCR with general eubacterial primers and subsequent hy-

bridization experiments with the amplified DNAs also yieldednegative results with the ¹tmbacter probe. The fact that thereare several nitrite oxidizers in other subdivisions of the Pro-teobacteria could readily expl~ our results, Three known au-totrophic nitrite-oxidizing bacteria are in the 8 subdivision, andone such organism is in the p subdivision, although a recentstudy has suggested that the phylogenetic placement of thegenus Xitrospira may need to be reconsidered 8!. The 8 sub-division nitrite-oxidizing bacteria were isolated from marineenvironments, but the salinity of the water from which thesample of iVitrospina gracilis was isolated was low �2.870 ppt!�7!. Ehrich et aL 8! recently isolated a new obligately chemo-lithoautotrophic nitrite-oxidizing species, ¹trospira moscovien-sis, which was cultured in freshwater media. Since few strainsof the other nitrite-oxidizing bacteria have been cultivated, itmay be premature to design probes based on only a few iso-lates.

There have been no definitive studies of the microbiology ofaquarium biofilters. Johnson and Sieburth �5! used scanningelectron microscopy to investigate nitrifying bacteria obtainedfrom the biological filters and waters of three aquacultureoperations one freshwater and two seawater!. These authorswere unable to detect bacteria with ¹trosomonas-like cyto-morphological features in actively nitrifying I'reshwater salmonculture systems. In addition, they could not find ¹trobacterwinogradskyi in any of the natural systems which they sampled,but these bacteria were found in subsequent enrichment cul-tures, These results are consistent with our results obtainedwith freshwater aquaria, in which no "classical" fVitrosomonasspecies could be detected.

To a certain extent, models of nitrification are dependent onthe known biochemical properties and pathways of the classicalnitrifiers. The data from our study indicate that the bacterialspecies responsible for nitrification in simple freshwater sys-tems remain unknown. It is likely that nitrification and theassociated nitrifying bacterium diversity in natural systems areeven more complex. Therefore, models which assume, in ageneral fashion, that lVitrosomonas spp, are the major nitrifiersmay have to be revised as novel species of nitrifying bacteriaare identified, isolated, and characterized and the biochemicalproperties of these species are determined. Molecular phylo-genetic methods, along with classical isolation and culturetechniques, afi of which are aimed at determining the respon-sible organisms and their physiological properties, should pro-vide a more complete understanding of biogeochemical pro-cesses mediated by nitrifying bacteria.

ACKNOWLEDGMENTS

We thank Ellen ICo, Quynh Lu, and Michefie Waugh for helpfulassistance. We also thank Julia Sears-Hartley, Melissa Lokken, andLes Wilson for performing the water chemistry analysis.

This work was supported in part by National Science Foundationgrants OCE92-18523 and OPP94-18442 to E.F.D. and by assistancefrom Aquaria, Inc., to TA.H.

REFEIKNCES1. ~ K L, W. Ludwig, and K.-H. Schletfer. 1995, Phsiogeaetie idrntili-

cation and in situ detection of individual microbial cells without cultivation.Mierobiol. Rev. 59:143-169.

2. American Tyye Cnlture Collection. 1992. Catalogue of bacteria and bacte-riophages, 18th ed. American Type Culture Collection, Rockville, Md.

3, Belser, L w., and F L Schmidt. 1978. Diversity in the ammonia-oxidizingnitrifier population of a soil. Appt Environ. Microbiot 36:SS4 � 5SS.

4. Bock, K., K-P. Knops, B. Ahlers, and B. Harms. 1992. Oxidation of inor-ganic nitrogen compounds as energy source, p. 414-430. In A. Balaws, H. G.Truper, M. Dworkin, W. Harper, and K,-H, Sehleifer ed.!, The prokaryotes:a handbook on the biology of bmeria. Ecophysiology, isolauoa, identifica-tion, applications, 2nd ed. Springer-verlag, New York.

2896 HOVANEC AND DELONG APPL. ENVIRON. MICROBIOL

5. Cssdgnetti, D., and T. C. Hollocher. 1984. Heterotrophic nitrificatiou amongdenitrifiers. Appl, Environ. MicrobioL 4'I:620-623.

6. DeGrsnge, V., and R. Bardin. 1995. Detection and counting of IAtrobacterpopulations in soil by PCR. Appl. Environ. Microbiol, 61;2093-2098.DeLong, E. F. 1992. Arcbsea in coastal marine eaviroaments. Proc. Natl.Acad. Sci. USA 89:5685-5689.Ehrich, S., D. Bebreas, E. Lebedeva, W, Ludwig, and E. Boch. 1995. A newobligately chemolithoautotrophic, nitrite-oxidizing bacterium, A'itrmpiramoscoviensis sp. nov. and its phylogenetic relationship. Arch. Microbial164:16-23.

9. Gibson, J�E. Stscbebrandt, L B. Zablen, R. Gaptn, and C. R. Woese. 1980A phylogenetic. analysis of the purple photosynthetic bacteria. Curr. Microbiol. 3;59M4,

10. Giovannoai, S. J�E. F. DeLong, G. J. Olsen, and N. R. Pace. 1988. Phylogenetic group-specific oligadeoxynucleotide probes for identification of single microbial cells. J. Bacteriol. 170:720-726.

11. Hall, G. 1986. Nitrification in lakes, p. 127 � 156. In J. L Prosser ed.!, Nitri-fication. IRL Press, Oxford.

12. Head, L M., W. D. Hiorns, T. M. Embley, A. J. McCarthy, and J. R.Saunders. 1993. The phylogeny of autotrophic ammonia-oxidizing bacteriaas determined by analysis of 16S ribosomal RNA gene-sequences. J. Geu.Microbiol. 13 9:1147-1153.

13. Hiorns, W. D., R. C. Hastings, L M. Head, A. J. McCarthy, J. IL Saunders,IL W. Picknp, aud G, IL Hall 1995. Amplification of 16S ribosomal RNAgenes of autotrophic ammonia-oxidizing bacteria demonstrates the ubiquityof nitrasospiras in the environment. Microbiology 141:2793 � 2800.Hugueain, J. E., and J. Colt. 1989. Design and operating guide for aqua-culture seawater systems. Elsevier, Amsterdam.

14

15, Johnson, P. W., and J. IIL Siebarth. 1976. In situ morphology of nitrifyinglike bacteria in aquaculture systems. Appl. Environ. MicrobioL 31;423-432

16. Kawai, A., M. Sugiyams, R. Shiozaki, snd L S~ 1971, Microbiological studies on the nitrogen cycle in aquatic environments. Ivlem. Res. InstFood Sci. Kyoto Univ. 32:7-15.Killhsm, K. 1986. Heteratropbic uitrificatiou, p. 117-126, In J. I. Prasser17 ed.!, Nitrificatiou. IRL Press, Oxford.

18. Kaeaea, J. G., and L A. Robertson. 1994. Cainbined nitrification-denitrification processes. FEMS MicrobioL Rev. 15:109-117.

19. lane, D. J. 1991. 16SI238 rRNA sequencing, p. 115 � 175. In E. Stackebrandtand M. Goodfellow ed.!, Nucleic acid techniques in bacterial systematics.Wiley-lnterscience, West Sussex, England

20. Mskhdtv B, I, N. Larsen, M. J. McCanghey, R. Overbeek, G. J. Olsen, K,Fogel, J, Blaady, and C. R. Woese. 1994. The Ribosomal Database ProjectNucleic Acids Res. 22i3485 � 3487.

22. Nejidat, A., and A. AbetioviA. 1994, Detection of btt'ttosomanas spp. by

21. McCaig, A. K, T. M. Embley, and J. I. Presser. 1994. Molecular analysis ofenrichment cultures of marine aininonia axidisers, FEMS MicrobioL Lett.12ih363-368.

polyinerase chain reaction. FEMS Microbial. Lett, 12ik191 � 194.23. Orso, S., bL Gouy, E. Navarro, aad P. Normand, 1994. Molecular phyloge-

netic analysis of Xitrobacter spp. Int. J. Syst. Bacteriol. 44;83-86.24, Painter, H.A. 1986. Nitrificatio in the treatment of sewage and waste-water,

p. 185-211. In J. I. Prosser ed.!, Nitrification. IRL Press, Oxford,V. Prosser, J. I. 1989. Autotrophic nitrificatian in bacteria. Adv. lVlicrob.

Physiol. 30;125-181.Q. Schimel, J, P., M. K. Fimstoae, and K. S. Killham. 1984. Identification of

heterotrophic nitrification in a sierran forest soiL Appl. Environ, Microbiol.4th 802-806,

27. Seewaldt, K, K.-H. Schleifer, F Bock, and E. Stsdiebraadt, 1982. The closephylogenetic relationship of ¹robacter aud Ifbodopsetutomonas patusois,Acta MicrobiaL 131:287-290.

28. Stahl, D. A�B, Flesher, H. R. Mansfield, and L Montgomery. 1988. Use ofphylogenetically based hybridization probes for studies of rurninal microbialecology. AppL Environ. Ivficrabial 5tk1079-1084.

29. Tate, IL L, HL 1977. Nitrification in histosols: a potential role for theheterotrophic nitrifier. Appl, Environ. Microbiol. 33:911-914.

30. Teske, A�E. Aim, J. M. Regan, S. Toze, B, E. Rittmaan, and D. A. Stahl.1994. Evolutionary relationships among ammonia- and nitrite-oxidizing bac-teria. J. Bacterial. 176:6623-6630.

31. Voytek, lVI. A., sad B. B, Wanl 1995. Detection of ammonium-oxidizingbacteria of the beta-subdass of the class Proteobactena in aquatic sampleswith the PCR. AppL Environ. Microbial. 61i1444-1450,

32. Wagaer, M., IL Amsnn, H. lammer, aad IC-IL Schleifer. 1993. Probingactivated sludge with oligouudeotides specific for proteobacteria: inade-quacy of culture-dependent methods for describing microbial communitystructure. AppL Environ. MicrobioL 59:1520-1525,

33. Wagaer, M�G. Rath, IL Amann, H, P, Koops, and IL-IL Schleifer. 1995. Insitu identification of ammonia-oxidizing bacteria. Syst. Appl. Microbiol. Ilk251-264.

34. Wagner, M., G. Rath, H. P. Koops, J. Flood, and IL Amass. In situ analysisof niuifyiug bacteria in sewage treatment plants, Water Sci. Tecbnol., inpress.

35, Ward, B. B. 1982. Oceanic distribution of anononium-oxidizing bacteriadetermined by immunoiluorescence. J. Mar. Res. 4ik1155 � 1172.

36. Wmd, B, IL 1986. Nitrification in marine environments, p. 157-184, In J. I.Prosser ed.!, Nitrification. IRL Press, Oxford.

37. Watson, S. W�aad J. IL Waterbury. 1971. Characteristics of two marinenitrite oxidizing bacteria, Ititrospina gracilis nov. gen. nov. sp. and ¹tococcusntobttis gen. nov. sp. nov. Arch. Mikrobiol. 77:203-230.

38. Wheaton, F. W., J, N, Hocbheimer, G. K Kaiser, M. J. Kronos, G. S. Iibey,and C. C. Easter. 1994. Nitrification filter principies, p. 101 � 126. In M. B.Timmons and T. M. Losardo ed.!, Aquaculture water reuse systems: engi-neering design and management. Elsevier, Amsterdam.

39. Woese, C, R. 1987. Bacterial evolution. Microbiol. Rev. 51:221-271.

Vol. 64, No. 1

Nitrospira-Like Bacteria Associated with Nitrite Oxidationin Freshwater Aquaria

TIMOTHY A. HOVANEC, ' * LANCE T. TAYLOR,'t ANDREW BLAKIS,AND EDWARD F, DELONG'g

Department of Ecology, Evolittion and Marine Biology, University of California, Santa Barbara,Santa Barbara, California 93106,' and Aquana Inc., IvIoorpark, California 93021

Received 4 September 1997/Accepted 27 October 1997

Oxidation of nitrite to nitrate in aquaria is typically attributed to bacteria belonging to the genus Witrobaeterwhich are members of the a subdivision of the class proteobacteria. In order to identify bacteria responsible fornitrite oxidation in aquaria, clone libraries of rRNA genes were developed from biofilms of several freshwateraquaria. Analysis of the rDNA libraries, along with results from denaturing gradient gei electrophoresis DGGE! on frequently sampled biofilms, indicated the presence of putative nitrite~xidizing bacteria closelyrelated to other members of the genus IVitrospira. Nucleic acid hybridization experiments with rRNA frombiofihns of freshwater aquaria demonstrated that Nitrospira-like rRNA comprised nearly 5% of the rRNAextracted from the biofilms during the establishment of nitrification. Nitrite-oxidizing bacteria belonging to thea subdivision of the class protesbaeteria e.g., Xitrobacter spp.! were not detected in these samples. Aquariawhich received a coinmercial preparation containing ItiitrISacter species did not show evidence of Witrobactergrowth and development but did develop substantial populations of IVitrasptra-like species. Time series analysisof rDNA phylotypes on aquaria biofihns by DGGE, combined with nitrite and nitrate analysis, showed acorrespondence between the appearance of/IIitrospira-like bacterial ribosomal DNA and the initiation of nitriteoxidation. In total, the data suggest that IIlitrobaeter winogradskyi and close relatives were not the dominantnitrite~xidizing bacteria in freshwater aquaria. Instead, nitrite oxidation in freshwater aquaria appeared to beinediated by bacteria closely related to Ãitrospira moscoviensis and Ãitrospira marina.

APPLIED AND ENvlRGNMENPAL MIcRoBloLQGY, Jan. 1998, p. 258 � 2640099-2240/98/$04.00+ 0Copyright CI 1998, American Society for Microbiolo@r

The oxidation of nitrite to nitrate by chemolithoautotrophicnitrite-oxidizing bacteria NOB! in fish culture systems, rang-iiig Rom home aquaria to commercial aquaculture systems, isan important process. The accumulation of high concentra-tions of nitrite, which is toxic to fish and other aquatic organ-isms, is prevented by active nitrite removal by nitrifying micro-organisms. Nitrite is formed in aquarium systems Rom theoxidation of ammonia, the principal nitrogenous waste of te-leosts, by autotrophic ammonia-oxidiziiig bacteria AOB!.Thus, closed aquatic filtration systems usually provide a solidsubstratum, which is termed a biological filter or biofilter, topromote the growth of AOB aiid NOB. A variety of materialscan form the substratum of a biofilter, ranging from gravel tospecially engineered molded plastics. Biofilters can be sub-merged in the fiow path of the filtration system or can belocated such that the water trickles or percolates through amedium situated in the atmosphere outside the aquarium,before fiowing back into the tank.

Traditionally, the bacteria responsible for the oxidation ofammonia aud nitrite in aquaria were considered to be Witro-somonas eumpaea aiid Nitrobacter winogradskyi or their closerelatives, respectively �7, 18!. However, there is some indica-tiou that both X europaea and ¹ Ivinogradskyi may uot bepredominant components of actively nitrifying freshwateraquaria 9!. Iu seawater aquaria, however, II/, europaea audclose relatives do appear to comprise a significant proportionof the total eubacterial community, but N. winogradskyi wasbelow detection limits 9!.

* Corresponding author, Mailing address: Aquaria Inc., 6100 Con-dor Dr., Moorpark, CA 93021. Phone 805! 529-1111. Fax 805! 529-3030. E-msih hovanec@lifesci.lscf.ucsb.edII.

t Present address: Monterey Bay Aquarium Research Institute, P.O.Box 628, 7700 Sandholdt Rd., Moss Landing, CA 95039.

Chemolithoautotrophic NOB are phylogeuetically diverse,occurring in several subdivisions of the class proteobacteria Fig. 1!. The most well-studied members of this group of or-ganisms i.c., /V. vvinogradskyi arid close relatives! belong to thea subdivision of the class Proteobacteria �6!. IVitrospina gracilisaud Xitrococcus mobilis, which were first isolated by Watsonaud Waterbury �6!, were determined to be members of the 8and y subdivisions of the class proteobacteria, respectively �4!.Another NOB, ¹trospira marina, is phylogenetically afiiliatedwith non-NOB such as Leptospirillum femooxtdans �, 14, 16!.Based on phylogenetic analysis of 16S rRNA sequences, Erlichet ai. �! proposed a new phylum within the domain Bacteriafor these organisms Fig. 1!. A newly discovered NOB from afreshwater environment a corroded iron pipe in a heatingsystem!, ¹trospira moscoviensis, was recently found to be phy-logeiietically related to M marina �!,

Whether in pure culture or on biofilters, NOB are slowlygrowing organisms with doubling times from 12 to 32 h �, 5,7!. Therefore, in newly set up aquaria, ammonia aud nitrite canreach concentrations toxic to fish before a sufficien biomass ofAOB aud NOB becomes established. To reduce the length oftime for establishment of NOB on biofilters, commercial prep-arations of these organisms, in various forms of preservation,are available to seed the aquarium environment. These prep-arations range Rom essentially pure cultures of IIIitrobacterspecies to mixed cultures of autotrophic AOB and NOB or-gaiusms and to products which combine autotrophic nitrifyingbacteria with various species of heterotrophic bacteria. Paststudies have generally shown these mixes to be Ineffectual buthave not elucidated specific reasons for their poor perfor-mance �, 15!.

In this study, we observed that ¹trospint-IIke species ratherthan Nitrobacter species appeared responsible for oxidationof nitrite to nitrate in freshwater aquaria. A combination of

IrlITROSPIRrf-LIKE BACTERIA IN AQUARIA 259Vot 64, 1998

Agrobactenum tumefaciens

trosomonas europae~Amrnoirioox idizing

Irlrtrosomonas eutropha¹trosomonas ureae

uilriie-ox idtdngbacteria

Pseudomonas testosteroni

A lteromonas haloplanlai s

scheri chia coliMug netob

Leptospirinas aeruginosa

85-a-A-22 tatget group

a-A-19 target group

NitrOSPiragIOUP

'Nitrite-oxidizingbaucis

FIG. 1. Phylogenetic relationships of autotrophic NOB in the a subdivision of the class Piateobecterr'ri and the Irtirrospirrr group. Clone 710-9, an rDNA doneoriginating &om aquaria with active NOB populations, is most simfiar to NOB of the ¹trospiro group. The specificities of two oligonudeotide probes designed forIrttoorrtrire spp. are indicated by the boxed sections.

methods was used to investigate concurrently the appear-ance of NOB on biofilters and the oxidation of nitrite to ni-trate. Oligonucleotide probes, which target Witrospira and closerelatives, were developed and used to quantify this group atdifferent times during the establishment of nitrification. Dena-turing gradient gel electrophoresis DGGE! was used to mon-itor the appearance of ¹tmspira-like bacteria during the onsetof nitrite oxidation. The effectiveness of a commercial mix ofAOB and NOB was also evaluated.

MATERIALS AND MEITIODS

TABLE 1. The nucleotide sequences and positions of oligonucleotide probes for NOB

Wash Nonrarget bacteriaTx 'C!' temp Targeted grarip with exact match

'C! to probe sequence

Position nucleotides!

Base sequence�' ta 3'!Probe

664-F85 CAC CGG GAA TI'C CGC GCI' CCI' C 63.0 60.0 IV. moscoviensis, trt. marina, Noneaud 710-9 done

435 � 454 TCC ATC TTC CCI' CCC GAA AA 5S.5 56.0 IV. moscoviensis 710-9 clone None

S-G-Ntspa-0685-a-A-22

S-"-Ntspa-0454-a-A-19

' Probe names designated by the standard proposed by Alm ei aL �!.s K cog numbering.' Tri, temperature at which 50% of the bound probe is released f'rom the homologous hybrid.

Nndeic acid sampbng and extraction. For rRNA extractions fram aquariumgravel, individual gmvel samples �0 g! were placed in a polypropylene tube andcovered with 2.5 ml of low-pH bufi'er �0 mM sodium acetate, 10 mM disodiumEDTA! and processed as previously described 9!, For DNA extraction, gravelsamples were resuspended in cell lysis buifer �0 mM EDTA, 50 mM Tris-HCI,0,75 M sucrose! and processed as described previously 9!. Runples were storedat � 20'C untB extraction.

DNA was quantified by Hoechst type 33258 dye binding and fiuorometry DynaQuant 200; Hoefer Pharmacia Biotech, Inc., San Francisco, CalK!. rRNAwas quantified by measuring Azen Lambda 3B; Perkin Elmer!, assuming that 1Azen U corresponds to 40 p,g of RNA per ml.

Clone libraries of PCR-amplified rRNA genes. Clone libraries were derived&am nucleic acid extracts of aquarium samples. Bacteria rRNA gene fragmentswere amplified with the primers S-D-Baci-0011-a-S-17 8f; GTT TGA TCC TGGCIC AG! and 1492r eubacterial; GGT TAC CIT GTT ACG ACT T! orS-'-Univ-0519-a-A-18 �19r; GWA TTA CCG CGG CKG CI'G!. PCR condi-tions, cycle parameters, and reaction components were as previously described�!. PCR products were evaluated by agarose gel elecrrophoresis. PCR fragments

were cloned with a TA cloning kit Invitrogen, Carlsbad, Calif.!, as previouslydescribed �!.

DGGE analysis and profiling. For DGGE analysis, ribosomal DNA rDNA!I'ragments were amplified with the forward primer 35Sf eubacrerial; CCT ACGGGA GGC AGC AG! with a 40-bp GC clamp an the 5' end as described byMurray et al. �1! and the reverse primer S-*-Univ-0519-a-A-18 �19r; GWATTA CCG CGG CKG CTG!. PCR was performed on a Siratagene RobocyclerGradient 96 La Jofia, Calif.! with the manufacturer's reagents. PCR conditionsincluded a hot start SIPC! and a touchdown procedure �1!. Initial denaiuraiionat 94'C for 3 min was followed by a denaturation at 94'C far 1 min, a tonchdawnannealing &om 65 to 55'C for 1 min and 29 s rhe annealing time during thetouchdown increased by 1,4 s per cycle!, and primer extension at 72'C for 56 s the extension time was increased 1.4 s per cycle!. The final temperature seriesof the above thermal cycle was repeated for 20 total cycles, followed by a finalextension at 72'C for 5 min. Amplicons were examined by agarose gel electio-phoresis.

DGGE was performed with a Bio-Rad D-GENE System Bio.Rad Laborato-ries, Hercules, Calif.!. AB gels were S.5% acrylanude-4iis with Bio-Rad reagents D-GENE Electrophoresis Reagent kit!. Gel gradients were poured with Bio-Rad reagents D-GENE Electrophoresis Reagent kit! with a denaturing gradientof 20 to 60% where 100% denaturant is a mixture of 40% deionized formamideand 7 M wea! aud the Bio-Rad gradient delivery system model 475; Bio-Rad!.AB gels were run at 200 V for 6 h. The gels were visualized in one of iwo ways,For isualization and recovery of discrete DNA bands, the gels were first stainedfor 10 min in 250 ml of 1X Tris-acetaie-EDTA TAE! buffer, in which 100 p.l ofethidium bromide I mg/ml! was added, and then were washed for 10 min in 1xTAE buffer. For documentafion proposes, some gels were stained in VistraGreen diluted 1:10,000! Molecular Dynamics, Sunnyvale, Calif.! for 20 min,followed by a 20-min wash in 1>< TAE brdfer, and then were scanned with aFluorImager SI Molecular Dynamics!.

Individual bands were excised froin the DGGE gels with alcohol-sterilizedscalpeb. Extracbon of DNA &are the gel foBowed the methods of Ferris ei al.

260 HOVANEC ET AL. APPL. ENVIRON. MICROBIOL,

TABLE 2. Similarity ranking for clone 710-9 isolated fromfreshwater aquaria and members of the iVitrospiru group

% Similarity to rDNA of:

rDNA soiirce

710-9 sequenceN. rnoscoviensis 96.1jV. marina 87.4 87.6Lefrtospirillttrn sp, 79.9 80.3L. ferrooxidans 78.1 78:41lfagnetobacrerittrn bavaricutn 78.2 77.9

80.277.9 91.079.7 78.3 77.0

Positions 24 to 1284 of 710-9 E, colt numbering!.

8!. The excised band was placed in a sterile 2-mi screw-cap tube with 500 p.l ofsterile deionized water. The tubes were half filled with glass beads catalog no.11079-101; BioSpec Products, Inc., Bartlesvifie, Okla,! and placed in a mechan-ical bead beater Mini-beadbeater-8; BioSpec Products! for 3 min at the highestsetting. The processed DNA remained in the tubes at 4'C overnight. Afterovernight storage, the tubes were centrifuged at 3,200 X g for 8 min at 4'C toconcentrate the gel I'ragments. The supernatant was transferred to a clean Ep-pendorf tube.

To check the extraction eificiency, the supernatant was reamplified with theDGGE primers and r~ed by DGGE. An extraction was considered accept-able if it yielded a single band in DGGE analysis which comigrated with theoriginal DGGE band in the mixed population sample.

Oligonucleotide probe development aud hybridixation procedures. Two oligo-nucleotide probes were designed which specifically hybridize with tV. marino, N.mnvcoviertsts, and the Witrrrtpiru-like rRNA gene sequence isolated in this studykern biofilters. One probe S-G-Ntspa-0685-a-A-22! targets the biofilter-derivedtV trnqriru-Bke bacterium and both H. marina and tV. moscovicnsis. The secondprobe S-*-Ntspa-0454-a-A-19! targets the biofilter-derived tVitrospiru-like bac-terium and its closest cultivated relative, N moscovteitsis Fig. 1!. Probe matcheswere initially screened by BLAST �! and CHECK PROBE �0!. The probeswere synthesized by Operon Tech, Inc. Alameda, Calif.!. The nudeotide se-quences and positions of the probes are shown in Table 1.

Since no pure rRNA of the biofilter-derived ¹tnupim-like bacterium is yetavailable, in vitro-transcribed 16S rRNA was used as a template for temperatureof dissociation Ts! determinations and as a control in hybridization experiments.In vitro-transcribed 16S rRNA was synthesized as described by Polz and Ca-vanaugh �2!.

The Tss of the ohgouudeotide probes were determined by measuring theamounts of probe eluted over a series of increasing wash temperatures �3!. Forthese tests, 200 ng of template was immobilized on a nylon membrane Hy-bond-N; Arnersham! aud hybridized overnight at 45'C with szP-labefied probe.After hybridization, the membrane was washed at rooin temperature in 1 x SET�50 mM NaCl, 1 mM EDTA, 20 rnM Tris; pH 7.8!-1% sodium dodecyl suli'ate SDS! for 30 min ou a shaker table. Individual IBter strips were then placed in a0.5-m! Eppendorf tube containing 500 trJ of I X SET-1% SDS preheated to theinitial test teruperature. The Eppendorf tubes were placed in a thermal cycler Perkin-Elmer! and incubated for 30 min, The membrane was transferred to anew Eppendorf tube containing 1X SKI' � 1% SDS, and the temperature wasincreased and maintained at the elevated temperature for 30 min. After eachwash, the wash bulfer was transferred to a scinSlation vial containiug 3 ml ofscintillation cocktail Liquiscint; National Diagnostics, Atlanta, Ga.! and wasmixed, and radioactivity was quantified by liquid scintillation counting. Eachprofile was performed in duplicate.

rRNA from aquaria was slot blotted and quantified with nucleic acid probesdeveloped in this and an earlier study 9! rinder conditions previously described 9!. The methods for determining the relative amounts of rRNA-specific hybrid-ization signal from each probe were the saine as those previously described 9!.

Sequendng. Sequencing of SSU r DNA excised t'rom DGGE gels or dunes wasperformed directly with Sequenase 2,0 U.S. Biochemicals, Cleveland, Ohio!.

Experimental aquarium systems. Three sets of experiments in aquaria wererun to i! study the establishment of nitrifying bacteria and ii! determine theeifect of a bacterial additive, New aquaria, filter systems, and gravel were used foreach test. Samples of aquariuin water for the three tests were analyzed forammonia gas dfitusion membrane method!, nitrite azo dye method!, and ni-trate cadnuum reduction-azo dye method! by fiow injection analytss as previ-ously described 9!.

i! Bacterial additive test Six afi-glass aquaria were established with an airliftundergravel filtration system model KF720; Neptune Products, Mootpsrk, Cal-

if.! in a temperature-controfied laboratory mean air temperature, 26.0 + Id'C!.The aquaria were covered with glass lids but were not ifiuminated other than byroom ceiling lights which were on a 14- and 10-h light aud dark cyde, respec-tively. A 6.8-kg amount of natural aquarium gravel Kaytee Products, iivrindale,Calif.! was placed on top of the filtration plate. A 30-liter volume of city tapwater, passed through activated carbon, was added to each aquarium. Filtered airwas supplied to each aquarium from a conunon air source. Six fish Dartiooetfuifrirtnotus! were placed in each aquarium and fed 0.5 g of fish feed Aquar-

, Kal Kau Foods, Vernon, Calif.! daily over two feedings. Three of the aquaria the treatment group! were each given doses of 8 ml of bacterial additive Cycle;Rolf C. Hagen Inc., Mansfield, Mass.! on the first day and once every 7 daysafterwards for an additional 3 weeks. The other three aquaria were the controlgroup and did uot receive an additive.

Two samples of 10 g of gravel were cofiected from each aquarium on a weeklybasis, and nucleic acids were extracted and analyzed as described above,

ii! Time of NOB appearance. Three aB-glass aquaria were established asdescribed above. A 34-liter sample of city tap water, which was passed throughactivated carbon, was added to each aquarium, which contained 4,53 kg of gravel.Initiafiy, 0.71 mmol of filter-sterilized �.2-p,m-pore-size filter! ammonium chlo-ride was added to each tank, fofiowed by an additional dosing of 5.0 mmol ofNHsCl on the fourth day. On days 10, 15, 18, 23, and 30, further anunoniaadditions of 8.9 mmol were made to each aquarium. During the test, a total of50.4 mmol of aminonia was added to each aquarium. Water samples werecofiected daily.

Two 10-g samples of gravel were collected from each aquarium daily for 33days. To one sample, 2 ml of lysis bufier was added aad the sample was hozeu � 20'C! until rDNA was extracted by previously described methods. rDNA wassubjected to DGGE after undergoing PCR with the primers and conditionsdescribed above. The other sample was preserved with 2 ml of bead beatingbufier.

iii! Time series. Three aquaria were set up as previously described with 4.53kg of gravel and were fified with 30 liters of city water which had been passedthrough activated carbon. The test was run for 138 days, during which the aquariawere individuafiy dosed with 8.9 mmol of filter-sterilized �.2 tim! ammonia asammonium chloride! on the first and second days of the test. From days 12 to 78of the test, further additions of 8.9 nunol of ammonia were done on averageevery 3 days. A total of 246 mmol of ammonia was added to each tank during thetest. The water was sampled three times a week for chemical analysis. Theaquaria were run for 80 days with freshwater, at which time the water wasswitched to seawater �2 ppt! by draining and refilling with water mixed withartificial sea salts Mariueland Commercial Aqiiariums, Moorpark, Calif.!, Afterthe switch, the testing couunued for an additional 57 days.

Nucleotide seqneaee accession ad The nucleotide sequence reported in thispaper for clone 710-9 has been deposited in the GenBank database underaccession no. AF035813. 45 so ss so ss ro ra so

Tesvrersturs t'c!

'rm o 30 ss sa ss so ss so ss 70 rs seT~ i'C!

FIG. 2, Results of the Ts experunents for the probes S-G-Ntspa-0685-a-A-22and S-*-Ntspa-0454-a-A-19, with the 50% Ts indicated by the vertical line. Cf,rRNA of Ar. manrta; ~ and 0, transcnbed RNA of clone 710-9.

Voi.. 64, 1998 MTROSPIRat-LIKE BACTERIA IN AQUARIA 261

Samples were taken 17 and 31 days after the aquarium estab-lishment and ammonia additions started. A third library wasconstructed from DNA extracted from the material of a com-mercial biofilter constructed of thermoplastic material modelCBW-I; Aquaria, Inc.!, This filter had been set up for 109 daysin a system with daily dosing of ammonium chloride.

The second approach used to monitor and identify nitrifyingmicroorganisms was DGGE. The DNA extracted from aquar-ium gravel samples taken during the establishment of nitrifi-cation was subjected to DGGE to produce a pattern of discretebands. The banding patterns were compared to each other andto band patterns produced by a mix of known nitrifiers. Uniquebands were excised from the gels and sequenced,

The sequences Irom the clone libraries and DGGE werecompared to bacterial sequences found in public databases BLAST [2] and RDP [10j!, Some clones, which showed aclose similarity to those of known nitrite-oxidizing organisms,were more completely sequenced,

Identificution of putative ¹trospirs-like NOB. Five sampleswere screened for NOB by either clone library development orDGGE. A total of 96 clones or excised bands were partiallysequenced. Of these, 11 were highly similar to members of the¹trospira group but none were similar to Mfrobacfer spp. Thepartial sequences were most highly similar to those of X. rpsa-

A � a e

4 .e ec

al e-p eac .peae.a 4-15 �! 5-G-Nlapa ppae-a-e-ap cl 5-'.Nlapa.ceca-a 4.19

FIG. 3. Spccifici ics of the oligouuclcofide probes targeting NOB of theMiaospira group sud the 710-9 done identified iu this study. Probe order waseubacterial probe S-D-Bact-033$-a-A-1$ A!, Mfucepipa-like NOB pcobe 8-G-Ntspa-0685-a-A-22 B!, sud Mbrospi pa-Ifke NOB probe S-4-Ntspa-0454-a-A-19 C!, with 5RNA, transcribed RNA trRNA!, or PCR-smpiified rDNA, in thefollowing arrangement: slot A-l, Comamoa as tcstostero fi slot A-2, Akaligeacseuueaphus; slot A-3, Akafigca cs faecali5; slot A-4, Comapaonas acidovoruns; slotA-S, M vipuagpucfskyi cDNA!; slot A-6, ¹uobscter agihs rDNA!; slot B-I, done710-9 rDNA!; slot B-2, clone 710-9 trRNA!; slot B-3, M marina rDNA!; slotB-4, M marina trRNA!; slot B-S, M gracNs; slot B-6, Sha vanaffa puuefacicns.See text for description of methods.

RESULTS

Isolation of putative NOB. Two approaches were taken toidentify NOB in aquarium samples, The first approach was todevelop clone libraries from gravel samples from an aquariumat several times during the establishment of nitrification,

TABLE 3, Results of probing rRNA extracted from aquarium biofilms with nucleic acid probes for NOB

Signal with the following oligouudeo ide probes"f:Daily

ammoniasmt

Aquariumenvironment'

Biofilmsubstrate

Samplelabel

Ammoniasource

NOBS-D-Beet-0338-

a-A-18 S-*-Nbac-1017- S-G-N spa-0685- S-'-Ntsps-0454-a-A-20 5-A-22 a-A-19

Type of aquarium water.Media from which the bacterial cells were extracted, Poiypp, polypropyiene.Fish, the aquarium had a fish population snd ammonia wss generated by the fish; NH4CI, the tank hsd no fish sud the ammonia wss from daily dosing with

smruonium chloride.4 Values in grams are the amounts of fish feed put into the aquarium each day; molar or miifimolsp values are the concentrations of ammonia added fo the aquarium

or system in which the biofilter was located each dsy.' +, signal detected by probe; �, no signal detected.iS-*-Nbsc-1017-3-A-20 wss originally called NBAC2 9! sud targeted M winognubhyi sud M agifis. Probe 8-G-Ntsps-0685-5-A-22 targeted M marina, M

a osc4avfcesis, end done 710-9. Probe 8-4-Ntsps -0454-s-A-19 targeted M a pascovicnsis aud clone 710-9.

710r7115T825T82SWDF 1036WDF 1036WDF1026WDF 1039WDF1038WDF 1035FLRT6EXPSBFWSW4FWSW6BC2-8BC2-10BC2-12BC2-13BC2-16BC2-4BC2-16s714r715rFWSW2FWSW3FWSWSFWSW9

FreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterFreshwaterSeawaterSeawaterSeawaterSeawaterSeawaterSeawater

PolyfiberPolyfiberPolyppGrave!PolyppCaravelPoiyppGravelSpongePolyppGravelPolyppPoiyppPolyppGravelGravelGravelGravelGravelGravelGravelPolyfiberPolyfiberPolyppPoiyppPolyppPoiypp

32.1 mM32.1 mM0.8 gO.S g3.2 g3.2 g2.0 g3.2 g2.0 g2.0 g2.0 g1.4 gSmM5mMSmMSmMSmMSmM5mM2.0 g2.0 g

714 mM714 mM

5mM5mMSmMSmM

NHGCINH,CIFishFishFishFishFishFishFishFishFishFishNH4CINH4CINH4CINH4CINH4CINH4CINH4CIFishFishNH4CINH4CINH4CINH4CINH4ClNH4CI

+ + ++ +

+ + +

+ + + + + + + + + + + + + + + + + + ++ + + + + + + + + + ++ + + + + + + + + + + + + + + +

262 HOVANEC ET AL. APPL. EKYIROyi. MICROBIOL.

12

10

6

Es 6e E 4

1 2.0

10.0

6.0

6.0

Z 4.0

2.0

0.0

12.00

10.00

6.00E

6.00

z 400

0.00 0 20 40 60 60 100 120 140Time days!

FIG. 4. Ammonia A!, nitrite B!, and nitrate C! chemistry for an aquai iumfram startup through 138 days. The saw-toothed pattern for ammonia is theresult of the increasing frequency of dosing with ammonium chloride as nitrifi-cation was being established. The water was switched ham heshwater to seawa-ter on day 80.

rina and N. moscoviensis data not shown!, The 16S rDNA ofa representative clone which contained the Nitrospira-likerDNA was fully sequenced, and a phylogenetic tree was in-ferred. Phylogenetic analysis indicated a high similarity be-tween this cloned rDNA �10-9! and meinbers of the Nitrospiragroup, N. mascoviensis and ¹ marina Fig. 1!. The rDNAcontained in clone 710-9 was 96.1% similar to that of N. mosco-viensis and 87.4% similar to that of N marina Table 2!.

Oligonucieotide probe specificity. Oligonucleotide probe se-quences, positions Escherichia coli numbering!, Tds, washtemperatures and target groups for the probes are indicated inTable 1, For probe S-"-Ntspa-0454-a-A-19, the Td was 58.5'C,while the T�was 63.0'C for the S-G-Ntspa-0685-a-A-22 probe Fig. 2!.

Slot blot experiments confirmed that the probe S-G-Ntspa-0685-a-A-22 was specific to the known NOB of the Nitrospiragroup, as well as to the clone 710-9 Fig, 3!. As predicted,probe S-'-Ntspa-0454-a-A-19 hybridized to clone 710-9, butnot ¹ maritur. Furthermore, experiments demonstrated thatneither probe hybridized with NOB which are members of thett or 8 subdivisions of the class Profeobacferia Fig. 3!.

Detection of NOB in aquaria. Table 3 summarizes the re-sults from the probing of several aquarium biofilms with theNOB probes. Probe S-G-Ntspa-0685-a-A-22 yielded a positivesignal with all freshwater and saltwater aquaria tested, Theprobe S-'-Ntspa-0454-a-A-19 detected Nitrospira-like bacteriain all freshwater aquaria, but not in all the saltwater aquaria Table 3!, There were no cases of positive detection by a probewhich targets cs proteobacterial Nitrobacfer species Table 3!.

Time series. The aminonia, nitrite, and nitrate values for arepresentative test aquarium dosed with ammoniuin chloridefor 138 days are shown in Fig. 4. The data show the expectedpattern for the establishment of nitrification in aquaria. Ini-tially, the concentration of ammonia increased and then de-creased to undetectable levels by day 12 the saw-toothed pat-tern. of the ammonia values is the result of the increasingfrequency of ammonia additions!, By day 12, the amount ofnitrite increased, reaching its inaximum value on day 22. Byday 38, the amount of nitrite was essentially 0 and that ofnitrate was steadily increasing Fig, 4!, The change from fresh-water to seawater at day 80 resulted in an immediate increasein the ainounts of ammonia and, subsequently, nitrite, It tooknearly 20 days for ammonia oxidation to become reestablished.Reestablishment of nitrite oxidation took approximately 40days.

A DGGE profile for selected days over the first 101 days forthis aquarium shows that the Niirospira-like rDNA sequenceappeared faintly on day 15, corresponding to the onset ofnitrite oxidation Fig, 5!, By day 22, the band corresponding tothe Niirospira-like rDNA sequence increased in relative inten-sity and remained intense over the next two sampling dates.After the switch to seawater, the relative intensity of theNitrospira-92ike band diminished, The general band pattern alsochanged qualitatively between freshwater and seawater sam-pling dates. The banding pattern for day 87 � days after theswitch! appeared to more closely resemble the pattern for day57 freshwater! than the pattern for day 101 seawater! Fig. 5!.

Time of Nitrospira-like bacterial appearance. The daily con-centrations of ammonia, nitrite, and nitrate over the first 33days after setup of a new aquarium are presented in Fig. 6. Thetrends were as expected, with ainmonia peaking about day 12.Nitrite values increased starting at day 12, peaked at day 21,and decreased to below detection limits by day 26. Nitratevalues steadily increased from about day 15 onwards. DGGEshowed that the band corresponding to clone 710-9, the puta-tive NOB, first appeared on day 12, with the relative intensityof the 710-9 band increasing daily based on relative fiuores-cence units of rDNA amplicons Fig, 7!,

A 8 c 0 E F G H I J

Time since aquarium stert-up days!

FIG, 5, DGGE time series profile from a biofilm of a freshwater aquariumduring the establishment of nitriiication. The aquarium water was switched toseawater on day 80. Lanes A, G, and f contain twa donee, including clone 710-9,a putative NOB showing close similarity ta the IVirrospira group. The bandcorresponding to this organism first appears with significant intensity an day 22.Lanes B, C, D, E, and F are sampling dates before the switch to seawater. LanesH and 1 are sampling dates after the switch to seawater. The water chemistry forvarious fonna af nitrogen in this aquarium is indicated in Fig. 4.

VoL. 64, 1998

1.20

1,00

> 060E

'e 0.600EE 0,40

0.20

0.00

1.20

1.00

0.60la

0.60

a0.40

Z0.20

Time since aquarium start-up days!

e e e 1I

4 ale e

e

DISCUSSION

0.00 0 5 10 15 20 25 30 35Time days!

PIG. 6. Inorganic nitrogen values for a newly established frcshwatcr aquar-ium dosed with ammonia chloride over 33 days. A! Ammonia O! values alongwith dates of ammonia additions X!; B! nitrite f3! and nitrate I! values forthe same aquarium. A DGGE profile of the nitrifying assemblage associated withthis aquarium is presented in Fig, 7.

Commercial additive. The addition of a commercial bacte-rial mixture which contained Nitrobacter sp., but not Nitrospi rasp., did not result in the detection of ¹trobacter species byoligonucleotide probe hybridization experiments Fig, 8!, How-ever, a band which comigrated with a control derived froinpure Nitrobacter DNA could be detected in the original com-mercial mixture by DGGE analysis data not shown!. ¹tro-spira-like rRNA was readily detected in the aquarium. ¹tro-spira group-specific probes indicated that the tank whichreceived the additive had a significantly greater percentage ofthe ¹trospira species rRNA Fig. 8!. By day 16, approximately5% of the eubacterial rRNA hybridized with the general ¹-trospira group-specific probe, compared to only 0.33% of theeubacterial rRNA in the tank which did not receive an additive Fig, 8!, By day 50, the values were 3,39 and 1.52% for theadditive and nonadditive aquaria, respectively Fig. 8!.

Nitrite concentrations in the two aquaria decreased as therelative percentages of ¹trospira-like rRNA increased. By day22, the nitrite value had reached a maximum in the tank whichreceived the additive. Nitrite concentrations reached maximain the nonadditive aquarium on about day 32. By day 38, thenitrite levels in both aquaria were essentially below our limitsof detection, and nitrate levels were equivalent in the treatedand nontreated aquaria Fig. 8!.

Our results from DGGE analysis, rRNA probing, and se-quencing generally indicate that Nitrospira-like bacteria are themost likely candidates responsible for nitrite oxidation in fresh-water aquaria. The combined use of molecular phylogenetic

%1TROSPIRA-LIKE BACTERIA IN AQUARIA 263

techniques and monitoring of water chemistry suggested a cor-respondence between changes in the biofilm microbial com-munity which coincided with the onset of ammonia and nitriteoxidation. The commencement of nitrite oxidation coincidedwith the appearance of the putative nitrite-oxidizing Nitrospira-like bacterium. The results lend support to the conclusion of anearlier study, which suggested that ts subdivision proteobacte-rial NOB Ãitrobacter types! were not major components ofnitrite oxidation bacterial populations in freshwater or marineaquaria 9!.

Results regarding the beneficial effects of the addition of abacterial additive containing Nitrobacter species were equivo-cal, While nitrite levels in treated aquaria decreased earlierthan those in nontreated aquaria, there was no evidence thatNitrobacter species were actively growing in these aquaria. It ispossible that the levels of Nitrobacter species were below theliinits of detection of our techniques. However, since ¹tro-spira-like bacteria were readily detected and that their estab-lishment coincided with nitrite oxidation we postulate thatNitrospira-like organisms, and not ¹trobacter species, are themajor nitrite oxidizers in the freshwater aquarium environ-ment, It is possible that the addition of bacterial inixturessupplies vitamins and other nutrients which generally stimulatethe growth of the nitrifying assemblages, fostering their growthand development and indirectly stitnulating nitrite oxidation.

In the present study, we identified ¹trospira-like putativeNOB by amplification of rDNA with general bacterial PCRprimers and DGGE analyses. We chose to use universal anddomain primers rather than group-specific primers, sinceprevious analysis suggested that nitrite oxidizers other than

0 2 5 5 10 11 12 13 14 15 15 15Time days!

PIG. 7. A! DGGE of select dates during the first 1S days after the startup ofa I'reshwater aquarium, during which time nitrification became established.Clone 710-9, a ¹rrurpint-like putative NOB, can be seen to appear starting atabout day 12 lane G!. B! Relative intensities of the band for clone 710-9 at eachsampling date. Associated water chemistry data for this aquarium are presentedin Fig. 6.

264 HOVANEC ET AL. APPL. ENVIRON. MICROBIOL

2.0O

5.50

6rs 1.00

0.50

ACKNOWLEDG1VIENTS0.00

4.003.503.00

5 2.50s 2,00B~ 1.50

1.000.500.00

0

Probes ior ~xidizioa bacteria

5.0

4.oa4 3.0sr 2.0Z+ LO

0.0Tank 4 Tank 16

Dsy 22Tank 4 Tank 16

Bey 50

10 20 30 40 50 60Tiros days!

~ S-'-ISSPa-lpipaa.A.IS Q SSSlbcprVCQIS-a.a-xs ~ S 'Srsar-Ioiro-A-2C

FIG. 8. Water chemistry data aad uudeic acid probe hybridization results fora freshwater aquarium during the first 57 days after sicstup. Nitrite A! andnitrate B! are for two tanks, i.c., tank 4 R!, which did noi receive a commercialbacterial ruixture, aud tank 16 �!, which received weekly additions of a com-mercial bacterial mixture for the first 4 weeks. Problems with aitratc analysisequipment resulted in ao data for days 18 through 40. C! Pcrccnt hybridization relative ta that of a eubactcrial probe, S-D-Bact-0338-a-A-18! to probes specificfor NOB. Probes S-G-Ntspa-0685-a-A-22 aud S-'-Ntspa-0454-a-A-19 target bti-trasprrs spp., while probe S-'-Nbac-1017-a-A-20 is for NOB of the a subdivisionof the class Prateabaerena.

Mtrobacter might be involved in mtrification in aquaria 9!.Combined monitoring of euvirorlmerital conditions waterchemistry! with bacterial assemblage analysis DGGE! allowedus tO deteCt a cOrreSpOndenCe between nitrite oxidation aitdWitrospira-like rRNA, By monitoring samples over time,changes in the microbial assemblage were evident. This ap-proach permitted a more focused efFort in the search for linksbetween environmental processes aiid the microbes which me-diate them.

When comparing biofilters, researchers in the past havebeen generally limited to assessing mainly water chemistrychanges, such as ammotua disappearanCe aitd nitrate appear-ance. The use of molecular probes for the relevant nitrifying

bacteria in different systems should provide a more detailedunderstanding of the interactiOn between the biology aitdchemistry of the systems. This in turn provides iitformatioiIrelevant to better filter design arid may allow the ejfects ofvarious conditions to be assessed with respect to their efFectsou the biology as well as the chemistry of the system.

We thank Ellen Ko, Quynh Lu, and Michelle Waugh for helpfulassistance and AlisoII Murray for assisting with the DGGE. We alsothank Julia Sears-Hartley, Mclissa Lokken, and Les Wilson for waterchemistry analysis.

This work was supported in part by National Science Foundationgrants OCE95-29804 and OPP94-18442 to E.F.D. and by assistancefrom Aquaria, Inc. to T.A.H.

REFERENCES1. Atm, E,W�D, B. Oerther, N. Larsen, D. A. Stahl, and I Rsskiu. 1996. The

oligonuclcotidc probe database. Appl. Environ. Mcrobiol. 62:3557-3559.2. Aitschul, S. F., W. Gish, W. Miller, E. W. Myers, aud D. J. Lipmsn, 1990,

Basic local alignment search tool. J. Mol. Biol. 215:403-410.3. Bsissr, L W., and E. L Schmidt. 1978. Diversity in the ammonia-oxidizing

uitriiier population of a soil. Appl. Environ, MicrabiaL 36:584 � 5S8.4. Bower, C. E., aud D. T. Turner, 1981. Accelerated nitrification in new

seawater culture systems: efectiveness of commercial additives aud seedmedia from established systems. Aquaculture 24;1-9.

5. Cariucci, A F�and D. H. StxicklansL 1968. The isolation, purification audsome kinetic studies of marine nitrifying bacteria Exp. Mar. BioL Ecol.2:156 � 166.

6. DcLong, E. F. 1992. Archaea in coastal marine environments. Proc. Natl.Acad. Sci. USA 89:5685-5689.

7. Ehrich, S., D. Behrsas, E. Lobodevs, W. Ludwig, snd E, Bock. 1995. Anew abligately chemolithoautotrophic, nitrite-oxidizing bacterium, Ãirra-spira rnaseaviensis sp. nov. aud its phylogeaetir. relationship. Arch. Mi-cr obiol. 164:16 � 23,

8. Ferris, M. J., G. Muyzer, snd D, M. Ward, 1996. Denaturing gradient gclelectrophoresis profiles of 16S rRNA-defiucd population inhabiting a hotspring microbial mat community. Appl. Environ. Micrabiol. 62:340-346.

9, Hovanec, T, A., snd E. F. DeLsng. 1996. Comparative analysis of nitrifyingbacteria associated with freshwater aud marine aquaria AppL Environ.MicrobioL 62:28S8-2896.

10. Maidak, B. L, N. Larson, M. J. McCaughsy, R. Ovesheokr G. J, Oissn, K,Fogsl, J. Blaudy, snd C, R, Wassa, 1994, The ribosomal database project.Nucleic Acids Res. 22:3485-3487.

11. Murray, A. L, J, T, Hoilibsugh, and C, Orrcgo. 1996. Phylogenetic compo-sitions of bacterioplanktou from two California estuaries campared by de-naturing gradient gcl clectropharesis of 16S rDNA fragments. Appl. Envi-ron. Microbiol. 62:2615 � 2620.

12. Polz, M. F., and C. M. Cavanaugh. 1997. A simple method for quantificatioof uncultured microorganisms in the environment based an in vitro tran-scription af 16S rRNA. Appl. Environ. Microbiol. 63:1028-1033.

13. Raskin, L�J. M. Stramley, B. E. Rittmauu, aud D. A. Stahl. 1994. Graup-spccific 16S rRNA hybridizatian probes to describe natural communities ofmcthaaogcus. Appl. Environ, Microbiol, 60;1232 � 1240,

14. Tesks, A�E, Aim, J, M. Rcgan, S. Tore, B. E. Rittmanu, snd D. A. StahL1994, Evolutionary relationships among ammonia- aad nitrite-oxidizing bac-teria, J, Bacteriol. 176:6623-6630.

15. Tinunermans, J. A., and IL GsrasiL 1990. Observations sur l'utilisation enctangs de suspensions bactcrieunes du commerce. Bull. Fr. Pbche Piscicult.31tk28 � 30.

16. Watson, S. W., aud J. JL Waterbury. 1971. Characteristics of two marinenitrite oxidizing bacteria, IVirrospina gnrcilis nav. gen. nav. sp. aud Ni arrcoccusmobilis nov. gen. nov. sp. Arrh. Mikrobiol. 77:203 � 230,

17. Wheaton, F. W. 1977. Aquacultural engineering. John Wiley rk Sons, lnc.New York, N,Y,

18. Wheaton, F, W�J, Hochheimer, snd G, E. Kaiser. 1991. Fixed film nitrif-icatio in filters for aquaculture, p. 272-303. In D. E Brune aud J. R. Tomasso cd.!, Aquaculture aud water quality. The World Aquaculture Society, BatonRouge, La,

264 HOVANEC ET AL APPL. ENVIRON. MICROBIOL.

1.50XEo 1.00

0.50

ACKNOWLEDGMENTS0.00

4.003 rtc3.00

8 250o 2.00g '1.50

1.000.500 JI0

0

rrraaaa lar Sarrkaeairsrrae eaaiarra

5.0

c40

'5 3.0

sr 2.0

m IS 1.0

10 20 30 40 50 60Time days!

~ a -raa~aaaWia fig See,~a.za ~ a. arear-iarr~aorr

0.0 Tank 4 Tank 16 Tank 4 Tank 16oay 22 Oay 50

FIG. 8. Water chemistry data and nudeic acid probe hybridization results fora Ireshwater aquarium during the first 57 days after startup. Nitrite A! andnitrate B! are for Iwo tanks, i.e., tank 4 8!, which did not receive a commercialbacterial mixture, and tank 16 �!, which received weekly additions of a com-mercial bacterial mixture far the firsi 4 weeks, Problems with nitrate analysisequipment resulted in no data for days 18 through 40. C! Percent hybridization relative to that of a eubacierial probe, S-D-Beet-0338-a-A-18! ta probes specificfor NOB. Probes S-G-Nispa-0685-a-A-22 and S-'-Ntspa-0454-a-A-19 target ¹-trospim spp., while probe S-'-Nbac-1017-a-A-20 is for NOB of the rz subdivisionof the dass Proteobaeteria.

lVitrobtzcter might be involved in nitrification in aquaria 9!.Combined monitoring of environmental conditions waterchemistry! with bacterial assemblage analysis DGGE! allowedus to detect a correspondence between nitrite OXidatiOn aridVitrospirtr-like TRNA. By mOiutOTiiIg SampleS Over time,changes in the microbial assemblage were evident, This ap-proach permitted a more focused effort irI the search for linksbetween euviroumeutaj processes and the microbes which me-diate them.

When comparing biofilters, researchers iu the past havebeen generally limited to assessing mainly water chemistrychanges, such as ammonia disappearance and nitrate appear-mice. The use of molecular probes for the relevant nitrifying

bacteria in different systems should provide a more detailedunderstanding of the interaction between the biology aiidchemistry of the systems. This in turn provides informationrelevant to better filter design aud may allow the effects ofvarious conditions to be assessed with respect to their effectson the biology as well as the chemistry of the system.

We thank Ellen Ko, Quynh Lu, and Michelle Waugh for helpfulassistance and Alison Murray for assisting with the DGGE. We alsothank Julia Sears-Hartley, Melissa Lokken, and Les Wilson for waterchemistry analysis

This work was supported in part by National Science Foundationgrants OCE95-29804 and OPP94-18442 to E.F.D. and by assistancefrom Aquaria, Inc. Io T.A.H.

REFERENCES1. Alm, E. W., D. B. Oerther, N. Lsrsen, D. A. Sbihl, and L Rasldn. 1996. The

oligonudeoiide probe database. AppL Environ. MicrobiaL 62:3557-3559.2, Altschal, S, F., W. Gish, W. MBIerr E. W. Myers, and D. J. Lipmau. 1990.

Basic local alignment search tool. J. MoL Biol. 215:403-410.3. Betser, I W., and E. L Schmidt. 1978. Diversity in the ammonia-oxidizing

nitrifier population of a sail. Appl. Environ. MlcrobioL 36;584-588.4. Bower, C. Kr and D. T. Turner. 1981. Accelerated niiiificaiion in new

seawater culture systems: eifectiveneaa of commercial additives and seedmedia I'ram established systems. Aquaculture 24:1 � 9.

5. Cavlaeci, A. F., aad D, IL Striddand. 1968. The isolation, puriiication audsome kinetic studies of umrine nitrifying bacteria. Exp. Mar. BioL Ecoi.2:156 � 166.

6, DeLoas, E, F. 1992. Archaea in coastal marine environments. Proc. Natl.Aced. Sci. USA 89:5685-5689.

7. Ehrich, S., D. Sehrens, E. Lebedeva, W. Ludwig, aad E. Bock, 1995. Anew obiigately chemolithoautotrophic, nitrite-oxidizing bacterium, ¹tro-spira moseoviensis sp. nov. and iis phylogenetic relationship. Arch, Mi-crobiol. 164:16-23.

8. Ferrhr, M. J�G. Mayaei; aad D. M. Ward. 1996. Denaturing gradient geleiecirophoresis proflles of 16S rRNA-defined population inhabiting a hotspring microbial mat community. Appl. Environ. Microbiol. 62:340 � 346.

9. Hovaner�T. A., aod E. F. DeLong. 1996. Comparative analysis of nitrifyingbacteria associated with freshwater and marine aquaria. AppL Environ.Microbioi. 62'2888-2896.

10. Maidah, IL L, N. Larsen, M. J. McCaaghey, R. Overhedr, G. J. Olsen, K.Fogel, J. Slandy, and C. R. Woese. 1994. The ribosomal database project.Nucleic Acids Res. 22i3485-3487.

11. IVIurray, A. L, J. T. Hollibangh, and C. Orregrc 1996, Phylogenetic coinpo-sitions of bacterioplankton from iwo California estuaries compared by de-naturing gradient gel elecirophoresiz of 16S rDNA fragments. Appl. Envi-ron. MicrobioL 62:2615-2620.

12. Polz, M. F., and C. M. Cavanaugh. 1997. A simple method for quantificationof uncultured microorganisms in the environment based on in vitro tran-scription of 16S rRNA. Appl. Environ. Microbiol. 63:1028-1033.

13. Rasldn, L, J. M. Stromley, B. E. Rittmsnn, snd D. A. StahL 1994. Group-specific 16S tRNA hybridization prabes to describe nsiiu ai communities ofmethsnogens. Appl. Environ. Microbial. 60:1232-1240.

14. Tasks, A., E. Alm, J. hrL Regaa, S, Toze, B, E, Rittmsnn, and D. A Stahl.1994. Evolutionary relationships among aminonis- and nitrite-oxidizing bac-teria. J. Bacteriol, 171k6623-6630.

15, Tinunermsns, J. A., and R. Gerard. 1990. Observations sur l'utilization enetangs de suspensions bacteriennes du commerce. Bull. Fr. Peche Piscicult.31rk28 � 30.

16. Watson, S. W., and J. IL Waterbary. 1971. Characteristics of iwo marinenitrite oxidizing bacteria, ¹troapina gracilis nov. gare nov. sp. and ¹trzrcoccrrsmobilis nov. gerc nov. sp. Arch. Mikrobiol. 77i203 � 230.

17. Wheaton, F. W. 1977. Aquscultursl engineering. John Wiley rk Sons, InnNew York, N.Y.

18. Wheaton, F. W., J. Hochheimer, sud G. E, Kaiser, 1991, Fixed film nitrifi-cation in filtets for aquaculture, p, 272 � 303. In D. E Brune and J. R. Tomasso ed.!, Aquaculture and water quality. The World Aquscrdture Society, BatonRouge, La.

Vol. 64, No. 5

Microbiology of a Nitrite-Oxidizing BioreactorPAUL C BURRELI�' JURG KELLER AND LINDA L BLACKALL

Advanced I4astetvater Management Centre, Bepartments of Microbiology andChemical Engineering, The University of Queensland,

Brisbane, 4072, Australia

Received 16 October 1997/Accepted 9 March 1998

The microbiology of the biomass from a nitrite-oxidizing sequencing batch reactor NOSBR! fed with aninorganic salts solution and. nitrite as the sole energy source that had been operating for 6 months wasinvestigated by microscopy, by culture-dependent methods, and by molecular biological methods, and the seedsludge that was used to inoculate the NOSBR was investigated by tnolecular biological methods. The NOSBRsludge comprised a complex and diverse microbial community containing gram-negative and gram-positiverods, cocci, and filaments. By culture-dependent methods i.e., micromanipulation and sample dilution andspread plate inoculation!, 16 heterotrophs � gram positive and 10 gram negative! were identified in theNOSBR sludge RC!, but no autotrophs were isolated. 16S ribosomal DNA clone libraries of the two microbialcommunities revealed that the seed sludge GC! comprised a complex microbial community dominated byProteobacterta �9%% beta subclass; 18%% gamma subclass! and high G+C gram-positive bacteria �0%!. Threeclones �%! were closely related to the autotrophic nitrite-oxidizer it/ttrospira moscoviensis. The NOSBR sludgewas overwhelmingly dominated by bacteria closely related to à moscoviensts 89%!. Two clone sequences weresimilar to those of the genus /IVitrobacter. Near-complete insert sequences of eight RC and one GC X rnosco-viensis clone were determined and phylogenetically analyzed. This is the first report of the presence of bacteriafrom the /I/ttrospira phylutn in wastewater treatment systems, and it is hypothesized that these bacteria are theunknown nitrite oxidizers in these processes.

MATERIALS AND MOODS

1878

APPLIED AND ENYIRoNMENTAL IvlicnoaioLooY, May 1998, p. 1878 � 18830099-2240/98/$04.00+ 0Copyright C> 1998, American Society for Microbiology

Nitrificatioii is the initial step in the removal of nitrogenouscompounds from wastewaters. It involves the two-step conver-sion of ammonia to nitrite ammonia oxidation! and nitrite tonitrate nitrite oxidation! �0!. Denitrification of the nitrate tonitrogenous gas removes the nitrogen from solution �8!. Ifnitrogen removal fails, the nitrogenous compounds passinginto waterways may cause a series of environmental and med-ical problems �!.

There are a range of autotrophic �1! and heterotrophicbacteria 8! capable of nitrification. Unlike heterotrophic bac-teria, autotrophs are dependent on this reaction to generateenergy for cell maintenance and growth. In wastewater treat-ment systems, autotrophs constitute only a small percentage ofthe mixed liquor microbial community, but they are responsi-ble for the bulk of nitrification �7, 18!.

In wastewater treatment systems, the genera Ãitrosomonas an ammonia oxidizer! and Nitrobacter a nitrite oxidizer! arethe two groups of autotrophs presumed to be responsible fornitrificatiort �1!. Although ammonia oxidizers have been in-tensively studied by the use of molecular methods �6, 27!, thenitrite oxidizers have not been similarly studied, In one studyof activated sludge flocs �5!, c/usters of /ttttrosomonas andXitrobacter spp. were adjacent to each other as revealed byfiuorescent in situ hybridization FISH! probing. However, irtother studies, it/itrobacter could not be detected, and it wasspeculated that other bacteria were likely responsible for ni-trite oxidation �2, 27!.

To investigate the identity of the nitrite oxidizers in waste-water treatment plants, a nitrite-oxidizing sequencing batchreactor NOSBR! was operated, After 6 months of operation

' Corresponding author. Mailing address: Department oi Microbi-ology, Advanced Wastewater Management Centre, The University ofQucensland, Brisbane 4072, Australia. Phone: 61 7 33651101. Fax; 617 33654620. E-mall: blackall@florcy.biosci.uq.edu.au.

of the NOSBR, the developed sludge RC! was investigated bymicroscopy, by culture-dependent methods, and by molecularbiological methods. In addition, the sludge used to inoculatethe NOSBR GC! was investigated by molecular biologicalmethods and compared with the NOSBR biomass.

Mixed liquor from the Merrimac Wastewater Treatment Plant at the GoldCoast, Queens and, Australia, was used as inoculum for the NOSBR. The Mer-rimac plant is a full-scale biological nutrient removal BNR! plant operating fornitrogen and phosphorus removaL Mixed liquor from the aerobic stage wascollected and brought to the laboratory on ice. A volume of 1 liter was used toinitiate the NOSBR, while further aliquots were stored at � 20'C.

Operation of NOSBR. The NOSBR was operated according to methods pre-viously reported �!. BrieRy, the reactor was a chemostat with an operatingvolume of 1 liter, and the reactor feed comprised the following per liter!; 400 mgof KNOt, 3.75 g of MgSOs 7HsO, 250 mg of CaCis 2HsO, 10 g of KHsPO4,10 g of KHsPO4, 200 mg of FeSO4 ~ 7HsO, and 20 g of NaHCOs pH 7.2!. Therewere four stages to each cycle and a hydraulic retention time of 12 h. The smgeswere i! feed, 500 ml of fresh medium for 30 min � to 0.5 h!; {il! aerobicreaction, 4,5 h �.5 to 5 h!; iii! settle, 40 min � to 5.7 h!; and iv! decant, 500ml of supernatant for 20 min �.7 to 6 h!. The total time per cycle was 6 h.

After the NOSBR was operated for a period of approximately 6 months, a10-ml grab sample of mixed liquor was removed from the reactor during themiddle of the aerobic reaction stage and used immediately for analyses,

Microscopy. Approximately 50 p,l of the NOSBR mixed liquor was Gramstained, and micrographs were taken with a Nihon Microphot FXA microscope.

CultmuAependent methods. The NOSBR sludge �00 p3! was washed twicewith phosphate-buttered saline PBS; 135 mM NaCI, 2.5 mM KCl, 10 mMNasHPOo 1.75 mM KsHPO4 [PH 7.5J!, resuspendcd in 400 al of PBS, andserially diluted to 10 ~. A volume of 50 Isl of each dilution was then spreadinoculated onto two types of agar media. These were Nutrient Ayu' NA; Oxoid,England! and autotrophic nitrite agarose ANA; composed of the reactor feed[see above] solidified with 10 g of agarose pet liter!. In addition, a range of thediluted sludge samples was btiegy sonicated, and individual cells were isolated bymicromanipulation �1! and inoculated onto ANA. Plates were then incubated at28'C until growth occwred. A range of colonies with diR'ercnt morphologiesgrew on the NA and ANA inoculated with sludge samples by spread inoculationand on ANA inoculated with mictomanipulated cells. These colonies were sub-cultured to ensure purity. The 16S ribosomal DNA rDNA! sequence was par-tially determined and analyzed for a range of these isolates by previously pub-lished methods �, 4!.

Var 64, 1998 NITRITE OXIDIZERS FROM SEWAGE 1879

Molecular biological methods. The total community DNAs i'rom the NOSBRsludge RC! and from the sludge used as inoculum for the NOSBR GC! wereisolated, and the 16S rDNAs were PCR ampgfied and doned.

DNA extraction. The biomass �00 p.l! was centrifuged at 12,000 x g for 5 min.The supernatant was discarded, and the pellet was resuspended in 500 p,l ofsaline-EDTA �50 mM NaCl, 100 mM EDTA [pH 8.0]!. A volume of 100 pl offreshly prepared 100-mg/ml lysozyme was added to the mixture and incubated at3TC for 1.5 h. The mixture was then subjected to four cydes of freezing andthawing at � 20 and 65'C, respertively, Following this, 100 p,l of 25% wt/vol!sodium dodecyl sulfate and 50 p,l of 2% wt/vol! proteinase K were added to themixture and the mixture was incubated at 60'C for 1.5 h. The DNA was recov-ered fram the tube by phenol-chloroform extraction �9!, The nudeic adds fromthe 0,5-ml aqueous phase were precipitated by adding 0.12 ml of sterile 3 Msodium acetate and 1 rnl of ice-cold 100% ethanol and incubating for 1 h at� 70'C. The DNA pellet was recovered by centrifuging the solution at 12,000 Xg for 20 min at 4'C, The pellet was washed by adding 500 pl of 70% ice-coldethanol and was recovered by centrifuging at 12,000 X g for 10 min at 4'C. Thepellet was then air dried, and the nucleic arids were dissolved in 100 p.l of sterilemilliQ-purified mQ! water. Residual RNA was removed fram the nudeic acidsolution by adding 3 pl of 10-mg/ml RNase and incubating at 3TC for 1 h. TheDNA was visualized by agarose gel electrophoresis �9!.

Amplification of the 16S rRNA genes �6S rDNA!. Amplification of the near-complete 16S rRNA genes tram the extrarted DNA was done by employing thebacterial conserved primers 27f and 1492r �4! in a PCR. The components of thePCR were 10 to 100 ng of DNA, 1 U of Tth Plus DNA polymerase Biotech,Perth, Australia!, 10 tsl of 10X reaction buffer Biotech!, 200 pM each! dATP,dCIP, dGTP, and dTTP deoxynucleoside triphasphates [dNTPs]!, and 1 pl of200-ng/pl each! primer in a final volume of 100 p,l made up with sterile mQwater. The reaction mixture was then overlaid with 80 p,l of mineral oil andplaced in a thermal cycler Feign-Elmer DNA Thermal Cyder 480!. A cyclingprogram of 30 cycles of 94'C for 60 s, 4FC for 60 s, and 72'C for 120 s with a finalextension of 1 cycle of 48'C for 60 s and 72'C for 300 s wss used, The ampiiconswere visualized by agarose gel electrophoresis and were purifie by the WizardPCR Cleanup Kit Promega, Sydney, Australia! according to the manufacturer'sinstructions.

Cloning of the 16S DNAs. Ainplieons were used immediately in a ligationreaction mixture comprising 1 pl of T4 DNA ligase � U/p,l!, 10x btdfer, 1 pl ofpGEM-T vector �0 ng!, 2 pl of amplicons �5 ng!, and 5 pl of sterile mQ water.All components except the amplicons were from the TA Cloning Kit lnvitrogen,Calif.!. Ligation occurred at 15'C for 16 h,

Ultracompetent Epicurian Coli XL2-Blue MRF' cells Stratagene, Sydney,Australia! were thawed on ice in preparatian for the transformation step. Avolume of 100 pl af thawed cells was gently placed in a chified 50-ml Falcon tube,and 1.7 p.l of [3-mercaptoethanol was added. The mixture wss incubated on icefor 10 min with regular gentle swirling. Then, 2 p.l of the ligation mixture wasadded to the cells, and the cells were incubated on ice for 30 min. A heat shockstep was done by immersing the Falcon tube in a 42'C water bath for exactly 30 s.Cells were then returned to ice for 2 min. A volume of 900 p.l of warm �2'C!sterile SOB �0 g of Bade Tryptone, 5 g af yeast extract, 0.5 g of NaCl per liter!was added to each tube of transformed cells. These were then shaken at 3TC for1 h.

A voluine of 25 pl of transformed cells was spread inoculated onto Luria-Bertani LB! agar plates containing ampicillin, 5-bromo-4-chloro-3-indolyl-[3-D-galactopyranoside, and isopropyl-P-a-thiogalactopyranoside LB Ampicillin/X-gal/1PTG!, �9! which were incubated at 3TC for 12 to 16 h and then at 4'C for1 h. Positive clones those containing 16S rDNA PCR inserts! appeared whiteand negative clones no inserts! were blue. Positive clones werc picked andpatched onto LB Ampicillin/X-gal/IPTG agar plates to ensure that the firstscreening was correct. Positive dones were picked, hamogenized into 300 pl of

i sterile 50% glycerol, and stored at � 20'C until required, These clones consti-tuted the done libraries.

Ampgfication of clone inserts. Stored dones fram the library were patchedonto LB Ampidllin plates from glycerol stocks and grown overnight at 3TC. Asterile tip from a F200 micropipettor was used to obtain a barely visible ainountof overnight growth, which was placed into a microcentrifuge tube containing 63p.l of sterile mQ water and 10 p.l of 10X reaction butfer, and the mixture wascovered with mineral oil. The tube was placed into the thermal cycler andincubated at 96'C for 10 min. Then, 200 p.M each! dNTP, 1 p.l of Tt/t Plus DNApolymerase, and 1 p.l of each of the plasmid primers SP6 and T7! �00 ng/pl;lnvitrogen! were added to each tube. PCR cycling and observation of ampliconswere performed as described above.

Restriction enzyme analysis REA! of clone inserts For the NOSBR RC!library, the amplicons from the SP6-T7 PCR hom individual dones were sub-jected to Hoelll Sigma, Sydney, Australia! digestion. Haeftl is a restrictionenzyme that recognizes and cuts the tetranudeotide sequence 5' GG-CC 3', i.e.,it is a "4-bp cutter." The digestion mixture consisted of 0.5 p,l of HaeIII enzyme�0 U/p,l!, 2 pl of NEB Buifer 2 Sigma!, 7 p.l of sterile mQ water, and 10 p.l ofamplicon, The reaction was carried out at 3TC for 3 h. The restriction-digestedfragments were visualized by electrophoresis in a 3% Tris-acetate-EDTA aga-rose gel �9! for 55 min at 80 V.

Clones containing inserts that produced identical restriction patterns were

grouped into operational taxonomic units OTUs!, and representatives of eachOTU were selected for insert sequencing and analysis.

Partial and near-txtmplete sequendng of clone iuserts. Amplicons fram theSP6-T7 PCR t'rom individual dones were purified with the Wizard PCR CleanupKit and sequenced with the ABI dideaxy sequencing kit ABI, Melbourne,Australia! according to the manufacturers' instructions and with primer 530f�4!. For some clones, near-complete insert sequence data were obtained. In thiscase, PCR of the inserts with the 27f and 1492r primers was employed. A rangeof baiterial conserved primers �7f, 357f, 530f, 927f, 1114f [4]! were used todetermine the sequences. PCR and sequencing were performed as describedabove.

Analysis of sequeuce data. The partial 16S rDNA sequences were comparedwith those on publidy accessible databases by using the program Basic LocalAlignment Search Tool BLAST [1]!. The sequences were also manually aligned,considering secondary structural constraints, with sequences froin members ofthe domain Bec/ena. Phylogenetic analysis of aligned data sets was carried out byusing the Phylageny Inference Package PHYLIP version 3.5! according topreviously published methods �!.

RESULTS

Microscopy. Over the 6-month period of exposure to a veryStttlP]e medium that favOred the grOWth of autatraphiC nitriteoxidizers, a diverse microbial community in terms of morphol-ogy cocci, rods, attd filaments! attd Gram stain reaction de-veloped.

Culture-dependent methods. Results for partial 16S rDNAsequences attd identities for 16 pure cultures of bacteria ob-tained from the NOSBR are shown in Table 1. A range ofbacteria were able to grow on the ANA medium, although ittsome cases, the growth took Up to 14 days. Itt addition, pro]i ]cgrowth of a range of bacteria was observed ott the NA. Clearly,the NOSBR Cotttmned heterotroPhs itt addition to autotroPhicnitrite oxidizers. A range of other bacteria itt addition to the 16reported were isolated. However, none were closely related toknown autotrophic nitrite-oxidizing bacteria.

Moleettlar biological methods. Inserts from a totd of 102C1OtteS frOm the RC library were examined by ~ attd theywere found to fall into 13 diKerettt OTUs Fig. I!. A total of 90clottes 88%! were grouped into OTU 1, while the retttainittg12 OTUs were each composed of individual clottes each ofthese OTUs was 1% of the total number of clottes!. Eachindividual OTU clone insert attd six representatives from OTU1 RC7, RC11, RC16, RC25, RC73, and RC99! were partiallysequenced. Results from BI&ST comparisons are given ittFig. 1. According to the BLAST results, the vast bulk of cloneinserts itt the RC clone library originated from bacteria whoseclosest relative is Nitrosptra moscoviensis. It is recognized thatBLAST analySiS iS a fairly cTude Way to align SequenCeS frOmclottes with those of specific bacterial genera or species, attd aselection of the ¹ moscoviensis-like c[ottes was analyzed [nmuch more detail. Also according to BLAST analysis, two oth-er clone inserts, RC44 OTU 3! and RC57 OTU 4! Fig. I!,most closely matched sequences from the genera Nifmbacterattd Bradyrhizobittm, respectively. These genera along withthe genera Afipia attd Rhodopseudotttonas are very closely re-lated according to rRNA comparisons �6, 23!. Phylogeneticanalysis attd direct paitwise comparisons of clottes RC44 andRC57 with their closest relatives did trot clearly align them withany one of these genera. However, the closest matches fromBLAST are given ift Fig. 1.

Inserts from a total of 77 clottes from the GC library werepartially sequenced with the 530f primer attd analyzed. Thegroups to which these clone inserts were a8iliated are shown ittTable 2. The majority of the clone sequences grouped withthe proteobacterial phylum �6%!, while 4% � clottes; GC3,GC86, attd GC109! grouped with the phylum ¹trospira. Thesequences of GC3 attd GC86 were 99% similar, while the se-

1880 BURRELL ET AL. APPL. ENYIRQN. MIGRoaioL.

TABLE 1. Results for isolates obtained from the NOSBR by either sample dilution and spread plate inoculation ormicromauipulation of individual cells to ANA

Information from BLAST comparisonGram stain aud ceII

morphologyIsolate'

No. of uucleotidcscompared

% Similarity withclosest match

Closest match

NA, Nutricut Agar; ANA, autotrophic nitrite agurose; S, spread plate inoculation; M, mieromuuipulatiou.b ND, uot described.

quence similarities between clone GC109 and clones GC3 andGC86 were 91.4 and 92.6%, respectively.

Analysis of Nitrospira clones. Near-complete insert se-quences were determined for eight RC clones seven fromOTU 1 [RC7, RC11, RC14, RC19, RC25, RC73, and RC99] andthe one from OTU 2 [RC90]! Fig. 1!, one of the three GC Ni-trospira clones GC86!, and four clones SBR1015, SBR1024,SBR2016, and SBR2046! phylogenetically grouped in the Ni-trospira phylum and from a clone library prepared by Bond etal. �!. The data were phylogenetically analyzed as shown inFig. 2. A similarity matrix of the 13 clone insert sequences andall those from N. moscoviensis and Nitrospira marina is shownin Table 3.

NOSBR has excellent nitrification capacities �!, and we in-vestigated its inicrobial community structure.

Microscopy and culture<ependent methods. The commu-nity is composed of complex morphological types and stiflretains the floccular nature of activated sludge. We were ableto isolate a range of heterotrophs on both ANA and NA byclassical sample dilution and spread plate inoculation and bymicromanipulation, The occurrence in the NOSBR of hetero-

1400 bp

DISCUSSION

Our goal was to discover the possible nitrite-oxidizing mi-croorganisms in wastewater treatment systems, since Wagneret al. �7! had unequivocally shown that Nitrobacter was notfound in sludges by FISH probing. Until then, wastewatertreatment personnel had presumed that Nitrobacter was thedominant nitrite oxidizer because it was commonly isolatedfrom sludges. Additional support for this notion came fromMobarry et al, �5! who used FISH to observe clusters of¹trobacter, closely juxtaposed with clusters of ¹trosomonas, inactivated sludge and biofilm samples. However, by quantitativemethods of rRNA extraction and slot blot hybridization, it wasconcluded that the contribution of Nitrobacter to nitrificationwas minor {15!.

Nitrobacter n subclass of the class proteobacteria! can growheterotrophically, while the remaining known nitrite oxidizers,Nitrospina 8 subclass!, ¹rococctts y subclass!, and Nitrospira Nitrospira phylum!, are unable to grow heterotrophically 9!,To preclude the selective advantage that ¹trobacter may havegained from heterotrophic growth, we employed strategies thatselected chemoautotrophic nitrite oxidizers. We attempted tosignificantly narrow the microbial cominunity from a complexmixture with multiple functions carbon, nitrogen, and phos-phorus removal from domestic wastewater! to a single func-tion autotrophic nitrite oxidation! of reduced diversity. Our

330 bp

� � � ~ 50bpc w

Ji

4I

s

0 g v c l5

F. i- i0 0

e w eOI w Vlie

ueCJ 0

KO 8

s cCV 4

Cl

0

ClCi

44ls L

s 0

'/e O tJCI

o %I0 gs

a L 92 .Che

0

CO

a W g

Z

u a

s C

FIG. 1. Diagrammatic representation of the banding profiles of the 13 OTUsin the RC clone hbraiy and their closest matches by BLAST comparisons withpartial 16S rDNA sequencing of inserts.

1-NA-S2-NA-S3-NA-S4-NA-S5-NA-S6-NA-S7-NA-S8-ANA-S9-ANA-S10-ANA-S11-ANA-S12-ANA-S13-ANA-S14-ANA-M15-ANA-M16-ANA-M

Acinetobacter sp.Bacillus firmusPseudomonas mendocinaPseudomonas alcaligenesAcinetobacter sp.Acinetobacter sp.Bacterial sp.Rhodococcus sp.Rhodococcus rhodochrousRhodococcus sp.Mycobacterium fallaxStaphylococcus epidermidisParacoccus aminovoransUnidentified actinomyceteStenotrophomonas sp.Comamonas testosteroni

422422500380400425375420434218280381300328315365

9098969796

10099979892949596979697

Single, gram-negative rodsLarge, long, gram-positive rods; chainsSingle, paired, gram-negative rodsLong, thin, gram-negative rodsShort rods; grain negativeShort rods; gram negativeNDGram-positive filamentsShort, fat rods; gram positiveGram-negative rodsLong, thin, gram-negative rodsGram-positive tetradsMedium-length, gram-negative rodsNDNDND

NITRITE OXIDIZERS FROM SEWAGE 1881Vor 64, 1998

TABLE 2. Phyla from the domaiu Bacteria represented in theseed sludge GC! clone library determined by BLAST

comparisons of partial clone insert sequences

%%uo of clonelibraryPhylum

Proteo bacteriaAlpha subclassBeta subclassGamma subclassDelta subclass

High G+C gram positiveLow G+C gram positive.Flerc/bacter/Cytophaga/Bacteroides,Nitrospira.Plaoctoinycctales.Unaffiliated ........................

529184

1075499

Escherrchra «.olr M24835! gamma ptoieobaclena5510!robacrer leinograds4'i L3550+6! alpha pto[e<lbaclefjarobacrer lrarrrbargensi s L>550~2

delta proreobacreoa L35504! L35503!

bucrerurrn banancum X7! 838!OPi-2 isolate L22045!Slackebrandl clone X86772!llum sp. M79383!Slackebrandl clone X86770!um Jerroosirlnns X72852! ge Slackebrandl clone X86773!m manna L3550! !a nrarina X82559!rrr nroscovi ensrs X82558!

trii rros!rrra phylumR1024R1015

SBR2046C86 Y14644}5 OTU1; Y14639!9 OTU1; Y 14638!

8R2016C7 OTU1; Y14640!14 OTU!; Y 14637!99 OTU1; Y14643!C73 OTU1; Y14641C90 OTU2; Y14642!11 OTU1; Y14636!

Clone erene 1

Clone cleric t

FIG, 2. Evolutionary dis!ance tree of the frtitrospiru phylum and other known nitrite oxidizers in the domain Bacteria based on a comparative analysis of 1,030nucleoudes, Most boobrirnp values greater than 9298 Rom 100 resamplings for distance numbers above branches! and parsimonious numbers below branches!analyses are presented at the nodes, The ou group, Bactenrides/ragilis, is noi shown in the tree. The bar represents 0.1 estimated change per uudeotidc.

trophic nitrification and of other reactions such as aerobicdenitrification cannot be ruled out, but the paucity of organiccarbon in the reactor would slow such reactions. Organic car-bon can theoretically come from the extracellular polymersthat the bacteria in the fiocs are producing and from deadmicrobial cu lls. The possibility of heterotrophic nitrificationand aerobic denitrification is currently being further investi-gated. We were unsuccessful in isolating an autotrophic nitriteoxidizer. However, the isolation procedures used are perhapsnot likely to favor this since we employed only growth on solidinedia and other groups have einployed liquid media for theisolation of nitrite oxidizers 9!.

Molecular biological methods. Previous studies generatingclone libraries with nonselected activated sludges e.g., that ofBond et aL [5!! indicated that the diversity of the communitywould be too great to simplify by REA. Consequently, for theseed sludge library GC!, partial insert sequencing was imme-diately done rather than REA for grouping. As with othersludges, the diversity of the Merrimac sludge was significant,

As well, Qe proteobacterial phylum dominated the library,comprising 56% of clones, with the majority of these being ofthe beta subclass of the class Proteobacteria �9% of all Bacte-ria!, The next largest group was the high G+C gram-positivebacteria �0%!. These findings are similar to those from otherresearchers where bacteria of the beta subclass and/or highG+C gram-positive bacteria are dominant in BNR systems �,13, 24, 25!, We did not recover any Witrobacter clones in theGC library but did identify three clones �% of the library!most closely related to / /. moscoviensis.

We employed REA for the grouping of clones from theNOSBR sludge library RC! because we hypothesized that themicrobial diversity should be reduced in this clone library,Culture-dependent methods had not supported such a hypoth-esis, but it is well recognized that these methods are heavilybiased and the results obtained with them are unrepresentativeof the true microbial composition �3!, However, because of arange of biases in the methods, clone libraries are not consid-ered adequate for generating quantitative information aboutthe diversity of the microbial community from which the librarywas prepared �!. Nevertheless, REA proved extremely usefulin grouping clones in the RC library because, when 102 cloneswere examined, one grouping comprised 90 88% of the total!clones. Of the remaining 12 clones, 11 contained inserts orig-inating from different bacteria with five from the Fleribacter/Cytophaga/Bacteroides phylum and five others from the pro-teobacterial phylum, Two Kiter!bacter-like clones could not beunequivocally aligned with any genus, but more sequence datafrom these clones could clarify their atfiliation. None of theclone inserts came from gram-positive bacteria, but 6 of the 16isolates reported were gram positive. In addition, gram-posi-tive bacteria were microscopically observed in the sludge. Celllysis methods !nay not have been rigorous enough to lyse thegram-positive bacteria or the primers may have preferentiallybound to the non-gram-positive templates in the PCR. Bond etal. �! also found very few gram positives in two clone librariesfrom sludges, but Wagner et ak �5! hypothesize that gram

1882 BURRELL ET AL. APPL ENVIRON. MICROBIOL

TABLE 3. Similarity matrix showing the percent similarities among 16S rDNA sequences of V. moscovrensis, X marina, aud13 near-complete sequences of clone inserts obtained from biomass from a full-scale BNR activated sludge

plant or an NOSBR and clones for which the partial sequences had been previously reported

% Sequence similarity with species of strain noeStrain Species or clone accession no.!no, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

IV. moscoviensis X82558!SBR1024SBR1015GC86 Y14644!SBR2046RC25 Y14639!RC19 Y14638!SBR2016RC7 Y14640!KC14 Y14637!RC99 Y14643!RC11 Y14636!RC73 Y14641!RC90 Y14642!IV. marina X82559!IV. manna X35501!

1 2

3 4 5 6 7 89 10111213141516

96.396.1 99.696.1 99.695.8 99993.4 93.493.2 93,193.0 92.792.9 '93.1.92,8 93.092.7 92.992.6 92.892.2 92.592,1 92.188.7 88.288.0 88.0

99.499.4 99.293.6 93,6 93.193.0 93.2 92.792.8 92,6 92.493.2 92.9 92.893.1 93.1 92.793.0 93.0 92.693.0 92.9 92,592.6 92.6 92.192.3 92.2 91.888,3 88.3 87.888.2 88.1 87,7

9$,899.1 98.798.7 98.79$.7 98.99$.5 9L798.5 98.798.0 98.298.1 98.688.1 87.687.9 87.5

9L59L5 99298.4 99Z 99.698.4 99.0 99.5 99.797.9 9$.7 99.1 99.4 99.49LO 98.1 9L6 98.$98.8 99.087.2 87.2 87,1 87.1 87.1 86.5 86,687.2 87.2 87.1 87.1 87.1 86.5 86.6 99.9

GC86 was obtained from a BNR activated sludge plant. RC clones were obtained I'rom an NOSBR. Partial sequences for SBR dones are reported in reference5. Values in boldface indicate the two environmental Ihrrospira clone dades.

PCR experiments will be complemented with FISH studies toquantify the numbers of Nitrospira in nitrifying systems, How-ever, we will use our PCR test as a screening in advance ofFISH probing of sludges to show the presence of nitrospiras,Data f'rom studies on nitrification kinetics from the enhancednitrite-oxidizing culture in the NOSBR �! will be combinedwith numbers of nitrospiras from FISH probing and used inmathematical modelling,

ACKNOWLEDGMENTS

P.C.B. had an Australian Postgraduate Award, and the research wasfinancially supported by the CRC Waste Management & PollufionControl Ltd., a center established by the federal government of Aus-tralia.

positives could be responsible for phosphorus removal in BNRsystems because increases in this population, as determined byFISH probing, were correlated with initiation of phosphorusremoval in activated sludge systems.

The RC clone library was predominantly composed ofclones 89% from OTU 1 and OTU 2! with inserts originatingfrom bacteria whose closest relatives are in the Nitrospira phy-lum and are most similar to Nitrospira moscoviensis. Directpairwise sequence comparisons between sequences in the two¹trospira clone clades see Fig. 2! showed that clone clade 1 SBR1015, SBR1024, SBR2046, and GC86! had an average16S rDNA similarity value of 99.4% Table 3!, while for cloneclade 2 RC7, RC11, RC14, RC19, RC25, RC73, RC90, RC99,and SBR2016! this value was 98.7% Table 3!. The averagesequence similarity between the two clone clades was 92.8% Table 3!, while those between N, moscoviensis and clonedades 1 and 2 Fig. 2! were 96,1 and 92.8%, respectively. Thehighest comparative value between an RC clone sequence andN. moscoviensis was 93.4% for RC25 Table 3!. From thesequence data analysis, the two clone clades would likely rep-resent two separate species. This conclusion is drawn fromdiscussions by Stackebrandt and Goebel �2!, who note thatorganisms with rRNA sequence similarity values of less than97.5% most likely represent different species.

Conclusions. From the data presented in this paper, wesuggest that the unknown nitrite-oxidizing bacteria in activatedsludges belong in the Nitrospira phylum. In the meantime, bothWagner et al, �7! and Schramm et al. �0! have discoveredclones originating from Nitrospira in industrial activated sludgeand biofilm processes, respectively. Both the seed sludge fromthe Merrimac plant and the highly selected autotrophic nitri-fying bioreactor biomass contain these organisms. Clones withinserts originating from Nitrobacter were not recovered in theGC library, but two RC clones RC44 and RC57! were closelyrelated to this bacterium and its relatives. In the meantime, wehave prepared ¹trospira-specific primers and in preliminarystudies involving a PCR test data not shown! have positivelycorrelated the presence of Nitrospira with excellent nitrificationin full-scale activated sludge plants, In addition, in processeswhere nitrification is poor, these bacteria are absent. These

REPERENCES1. Altsehul, S, P., W. Gish, W. Mgler, E. W. hfyers, and fy. J. Lipman. 1990.

Basic local alignment search tool. J. Mol. BioL 215i403-410.2. Argaman, Y, 1991. Biological nutrient removal, p, 85-101. In A. M. Martin

ed,!, Biological degradation of wastes, Elsevier Applied Science, New York,N.Y.

3, BIackall, L 1 1994. Molecular identification of activated sludge foamingbacteria. Water Sci. Technol, 29:35-42.

4, Blachau, L L, E. M. Sevioiir, M. A. Cuaningham, IL J. Saviour, aad P.Hugenholtz. 1994. "Irticrortink parviceita" is a novel, deep branching memberof the actinomycetes subphylum. Syst. Appl. Microbiol, 17;513-518.

5. Bond, P. L, P. frngenholts, J. Keuer, and L L Blaekalt, 1995. Bacterialcommunity structures of phosphate-removing and non-phosphate-removingactivated sludges from sequendng batch reactors. AppL Environ. MicrobioL61: 1910 � 191 6.

6. Barrett, P., L L Blaekall, and J. Keuer, Characterisation of the bacterialconsortium involved in nitrite oxidation in activated sludge. Water Sei. Tech-nol., in press.

'7, Barrett, P. C., L I Biaehaa, aad J. Keller. Unpublished data.8. Castigaetti, Bo and T. C, Hollocher. 1984. Heterotrophic nitrification among

denitrifiers. Appl. Environ. Microbiol. 42:620 � 623.9. Ehrich, S., D. Behrens, E. Lehedeva, W. Ladvrig, and E. Bock. 1995. A new

obligately chemohthotrophic, nitrite-oxidizing bacterium, Kirrrrspira mosco-viensis sp. nov. and its phylogenetic relationship. Arch. Mierobioh 164:16 � 23.

10. Hailing-Sarensen, B. 1993. Biological nitriheation and denitrifieation, p.41-53. In B. Hulling-Sarensen snd S. E, Jorgensen ed.!, The removal ofnitrogen compounds from wastewater. Elsevier, Amsterdam, The Nether-lands,

11. Halhng-gorensen, B., and S. E, Jergensen ied.!. 1993. The removal ofnitrogen compounds from wastewater. Elsevier, Amsterdam, The Nether-lands.

VoL 64, 1998 NITRITE OXIDIZERS FROM SEWAGE 1883

12. Hovanec, T, A., and E, F. DeLong. 1996. Comparative analysis of nitrifyingbacteria associated with I'reshwater and marine aquaria. Appl. Environ.Microbiol. 62i2888 � 2896.

13. Kampfer, P., IL Erhart, C. Beirnfohr, J. Bohringer, 14L Wagner, and ILAmann. 1996, Characterization of bacterial communities I'rom activatedsludge � culture-dependent numerical identification versus in situ identifica-tion using group- aud genus-specific rRNA-targeted oligonucleotide probes.tvlicrob. EcoL 32:101-121.

14. Lane, D. J. 199], 16S/23S rRNA sequendng, p, 115 � 175. Iu E, Stackebrandtand M. Goodfellow ed.!, Nucleic acid techniques in bacterial systematics.Academic Press, Chichester, United Kingdom.

15. Mobarry, B. K., M. Wagner, V. Urbain, B. E. Rittmanu, and D. A. Stahl.1996. Phylogenetic probes for analyzing abundance and spatial organizationof nitrifying bacteria. AppL Environ. Ivlicrobiol. 62:2156-2162.

16. Orso, S., M. Gouy, E. Navarre, and P. Normand. 1994. Molecu!ar phyloge-netic analysis of Nionbacter spp. Int. J. Syst. Bacteriol. 44:83-S6.

17. Randall, C. W. 1992. Introduction and principles of biological nutrient re-moval, p. 7-84. In C. W. Randall ed.!, Design and retrofit of wastewatertreatment plants for biological nutrient removal. Techuomic PublishingCompany Inc., Lancaster, Pa

18. Robertson, L A., and J. G, Kuenen. 1991. Physiology of nitrifying anddenitrifying bacteria, p. IS9-199. Iu J. E. Rogers and W. B. Whitman ed.!,Microbial production and consumption of greenhouse gases: methane, ni-trogen oxides, and halomethanes, American Society for Microbiology, Wash-ington, D.C.

19. Samhruok, J., E. F. Fritsch, and T. Maniatis. 1989. Ivlolecular cloniug: a

laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Hhrbor, N.Y.

20. Sehranuu, A., D. De Beer, H. vaa dea Heavel, S. Ottengraf, aad R. Amann.1997, In situ structure/function studies in wastewater treatment systems.Presented at the Second International Conference on Microorganisms inActivated Sludge and Biofilm Processes, Berkeley, Calif.

21. Sherman, V. B. D. 1968. A new type of micromanipulator aud microforge.J. Geu. MicrobioL 54:287-297.

22. Staehebrandt, E., and B. BL Goehel. 1994. Taxonomic note: a place forDNA-DNA reassociation aud 16S rRNA sequence analysis in the presentspecies definition in bacteriology, Int, J. Syst, Bacteriol. 44846-849.

23. Teshe, A., E. Atm, J. M. Regau, S. Toze, B, E, Rittmann, and D, A, StahL1994, Evolutionary relationships among ammonia- and nitrite-oxidizing bac-teria. J. Bacteriol. 176:6623-6630.

24. Wagner, M., R. Amann, IL Lemmer, W. Manz, and K. H. Schleifer. 1994.Probing activated sludge with fluoresceutly labeled rRNA mrgeted oligonu-cleotides. Water Sci. Techuol. 29:15-23.

25. Wagner, M., R. Erhart, W. Mana, R. Aiaann, H. Lemmer, D. Wadi, audK.-H. Schleifer. 1994. Development of an rRNA-targeted oligonuclcotideprobe specific for the genus Aciuctohacrer and its application for in situmonitoring in activated sludge, AppL Environ, Microbiol. 60:792-800.

26, Wagner, M., G, Rath, R, Amann, H. P, Koops, and K. H. Schleifer. 1995. Insitu identification of ammonia-oxidizing bacteria. Syst. Appl. Microbioi. IIL251 � 264.

27. Wagner, M., G. Rath, IL-P. Koops, J. Flood, and R. Amaun. 1996. In situanalysis of nitrifying bacteria in sewage treatment plants. Water Sci. Technol.34:237 � 244.